Exécutez n'importe quel Skill dans Manus
en un clic

Exécutez n'importe quel Skill dans Manus en un clic

compare-calibration-rounds

Étoiles6

Forks1

Mis à jour7 juillet 2026 à 07:52

Compare A2MC calibration rounds (R1, R2, …) against each other and the validation targets — top-N best-achievable biomass per round, per-target Morris μ* sensitivity, P-pool / cross-regime overlays. Use when the user asks to "compare rounds", "update the multi-round figure", "refresh the R1-RN comparison", "which round is best", "top sensitive parameters per round", or to regenerate any `multiround_*` / cross-round figure after new cases complete. Codifies the multiround bundle pipeline + the footguns (screening contamination, case-name patterns, μ* ranking, param-set mismatch, partial-ensemble caveats).

Installation

Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.

Exécuter dans Manus

Source

jingtao-lbl

jingtao-lbl/A2MC-elm

Ouvrir le dépôt GitHub Voir les dépôts du créateur

Téléchargement

Exécuter dans Manus

Métiers associésSOC

Basé sur la classification professionnelle SOC

Biologistes, autresSciences de la vie, physiques et sociales·SOC 19-1029

SKILL.md

readonly

name	compare-calibration-rounds
description	Compare A2MC calibration rounds (R1, R2, …) against each other and the validation targets — top-N best-achievable biomass per round, per-target Morris μ* sensitivity, P-pool / cross-regime overlays. Use when the user asks to "compare rounds", "update the multi-round figure", "refresh the R1-RN comparison", "which round is best", "top sensitive parameters per round", or to regenerate any `multiround_` / cross-round figure after new cases complete. Codifies the multiround bundle pipeline + the footguns (screening contamination, case-name patterns, μ ranking, param-set mismatch, partial-ensemble caveats).
modes	{"requires_fates":true,"nutrient_pathway":"any","scope":["analysis"],"summary":"Cross-round comparison of FATES Morris ensembles; requires FATES calibration rounds. Targets from the case targets.yaml."}

Compare Calibration Rounds

Cross-round comparison of A2MC calibration rounds: best-achievable biomass per round vs the validation targets, and Morris μ* sensitivity per round. The canonical bundle lives under a site's analysis dir, use_cases/<site>/analysis/multiround_*/ (self-locating scripts; figures versioned beside them; the Kougarok reference instance is multiround_top50_sensitivity_20260606/). Read README.md there + the originating dev/ana log (findings) and the originating dev/ana log (procedure) first.

When to use

"Compare the rounds", "which round is best", "update/refresh the multi-round figure".
"Top sensitive parameters per round" / cross-round sensitivity.
Regenerating a multiround_* figure after new cases complete (e.g. a round's relaunched/added cases).
P-pool or cross-regime cross-round overlays (same bundle conventions).

The bundle

multiround_top50_sensitivity_20260606/
├── scripts/
│   ├── plot_multiround_top50_overlay.py        # biomass top-50 vs 6 targets (6 panels)
│   ├── plot_multiround_sensitivity_overlay.py  # Morris μ* top params per round (grouped bars)
│   ├── build_Y_from_monthly.py                 # R2..R5 Morris Y matrices from monthly NCs
│   └── build_Y_R1.py                           # R1 (138-param)
├── screening_top50/{R1..RN}/                   # screen_ensemble.py outputs per round
├── sensitivity/{R1..RN}_{leaf,fineroot}/       # morris μ* CSVs + per-round png
├── Y_matrices/                                 # Morris inputs (for sensitivity only)
└── multiround_*.png                            # the comparison figures

Also: tools/compare_rounds.py for quantitative round-vs-round deltas.

Pipeline (to refresh a figure after new cases)

Extract the new cases (all 3 phases if the combined/sensitivity figures are needed): production tools/extract_monthly_variables_FATES.py (TRANS) or tools/extract_and_plot_selected_cases.py (extract, any phase via --phase/--year-start/--year-end, reuses process_case).
Re-screen the round: phases/phase2_screening/screen_ensemble.py --data-dir <round extract dir> --top-n 100 --output-dir screening_top50/<RN> (source that round's config so A2MC_CASE_NAME_PATTERN is right). ⚠️ see footgun #1 — filter out-of-range cases first.
Biomass figure: python scripts/plot_multiround_top50_overlay.py (reads screening_top50/*/).
Sensitivity (only if refreshing μ*): rebuild Y (build_Y_from_monthly.py) → Morris (phases/phase1_exploration/morris_sensitivity_analysis.py) → plot_multiround_sensitivity_overlay.py.

Footguns (the load-bearing part)

#	Footgun	Guard
1	CROSS-round screening must cap the case number. `screen_ensemble.py` matches `PtCNPEn(\d+)<suffix>_TRANS` for any digits, so out-of-Morris-range experiment cases (e.g. H1 clumping #5001/5005/5006) sitting in a shared extract dir get swept in and can rank as "best". (This made the R5 biomass figure briefly show best #5001 instead of #1304, 2026-06-10.)	Pass `screen_ensemble.py --max-case-num <ensemble max>` (e.g. `4890`) for cross-round comparison — it drops out-of-range cases and prints how many. Verify `# Sets:` in `_results.txt` = expected. NOTE: leave `--max-case-num` UNSET for a single round's own whole-ensemble summary graph — those legitimately exceed the Morris size (e.g. R3 + reruns = 4941).
2	Read `_results.txt` Sim_ columns + `screening_result.json` `best_case_num`, NOT `screening_top50_indices.txt`. The indices file mislabels case numbers on partial ensembles and isn't always rewritten (can be stale).	The overlay scripts already do this; don't "trust" the indices file.
3	Morris CSV `rank` column is by `mu`, not `mu_star`.	Re-sort by `mu_star` (Morris importance) for selecting/ranking top params.
4	Sensitivity file prefix is unreliable. Older runs (R1–R4, 6/6) named both organs `morris_leafbiomass_` (the `--output-var` quirk); newer runs (R5, 6/10) correctly use `morris_finerootbiomass_` for fineroot. So a dir can hold either prefix. The directory `{Rn}_{organ}` is the source of truth for the organ; content is correct per dir.	Map organ by directory, never by filename prefix. Glob `morris_PFT{X}_2.csv` (matches both); ensure one file per PFT so `sorted()[-1]` is unambiguous.
5	Param-set mismatch across rounds. R1 used 138 params; R2+ used 162.	Align on shared names; show "no bar"/blank for params absent in a round; note it in the caption.
6	Partial-ensemble rounds. Screening/ranking is valid on a partial ensemble (dead cases get finite high cost), but Morris μ is NOT reliable* until the ensemble is reasonably complete.	Include a partial round in the biomass top-N figure (label it), but exclude it from the sensitivity overlay until a full Y-rebuild + Morris re-run.

Conventions

Filenames embed round + axis-mode + case count: R{N}_{TRANS,combined}_{count}cases_ensemble.png, multiround_top50_biomass_vs_6targets_R1-RN.png, multiround_top_sensitive_params_bars_R1-RN.png (per feedback_plot_filename_convention).
Biomass overlay bands: grey ±20% acceptance band + light-blue ±1 SD observation band (SD often exceeds the mean → floor the lower SD edge at 0). SD values from validation_targets_leafroot.txt.
Whole-ensemble (not top-N) cross-round plots use tools/plot_ensemble_cases.py + regen_ensemble_milestone_plot.sh (different from this bundle's top-50 overlay).
NERSC: scratch (symlink dirs, filtered sets) under $HOME/repo tmp/ only — never /tmp,/scratch.

Cross-references

Worked examples: the originating dev/ana log (procedure), 20260610b (refresh + H1 catch), the originating dev/ana log (findings).
Tools: phases/phase2_screening/screen_ensemble.py, phases/phase1_exploration/morris_sensitivity_analysis.py, tools/{compare_rounds,extract_and_plot_selected_cases,plot_ensemble_cases}.py.

Changelog

2026-06-17: ## Changelog convention adopted (see .claude/skills/README.md). Earlier history: git log + memory/dev_logs/.

Plus depuis ce dépôt

même dépôt

a2mc-init

jingtao-lbl/A2MC-elm

First-run interactive setup for the offline agent — the official "getting started" flow the FIRST time someone uses A2MC in a repo/site. Interview the user (is a2mc_config.sh set up, where is the E3SM/ELM-FATES checkout, FATES on/off, carbon-only vs nutrient-enabled PARTEH, ECA vs RD, site name/location, PFTs, calibration targets), verify the checkout + RAG milestone, then create and populate the use case (site config + validation targets + parameter list) and hand off to phase0-design. Use when the user says "set up A2MC", "first time using A2MC", "help me get started / onboard me to A2MC", "initialize a new site / use case", "configure A2MC for my site", "I want to calibrate a new site". DISTINCT from onboard-session (which resumes an ALREADY-configured setup).

2026-07-076

add-skill

jingtao-lbl/A2MC-elm

Scaffold + register (or retire + de-register) an A2MC skill so the registry steps are never half-done. Use when the user or agent wants to create a new capability — "add a skill", "create/scaffold a skill", "make this reusable as a skill", "distill X into a skill" — OR to remove one — "remove/retire/delete the X skill". Writes (or deletes) the SKILL.md, and registers (or de-registers) it across all THREE human-facing registries (the README 'Current skills' table, docs/a2mc_reference/skills_catalog.md, AND the AGENTS.md 'At a glance' table), runs the mechanical drift check, and stops for human review before commit. Human-gated.

2026-07-076

arm-hpc-monitoring

jingtao-lbl/A2MC-elm

Set up real-time monitoring of an active A2MC HPC ensemble or experiment on Perlmutter at session start (CLAUDE.md Rule

2026-07-076

calibration-log

jingtao-lbl/A2MC-elm

Log interactive calibration work for a site under use_cases/{site}/memory/logs/. Two log types — a PHASE log (a diagnosis, screening, experiment design/result, refinement — recorded the same way the autonomous agent does, via PhaseLogger, so both modes' logs synthesize together) or a free-form SESSION log (exploratory work that does not map to a specific phase). Use when the user wants to record calibration or exploration work — "log this", "log this phase / diagnosis / experiment", "log this calibration session", "write a session log", "record what I explored".

2026-07-076

inject-knowledge

jingtao-lbl/A2MC-elm

Inject a human-originated domain fact (a discovery, a parameter insight, a mechanism/relationship) into A2MC's curated knowledge so the calibration agent surfaces it — placing it correctly across the up-to-three channels (site discoveries.json, generic parameters.json, curated_relationships.yaml), validating, and rebuilding the graph. Use when the user says "add this finding/parameter knowledge to A2MC's AI", "inject the X discovery into the KB", "make the agent aware of X", "author a curated relationship for X", or wants a manuscript/literature insight to reach the reasoning pipeline. This is the HUMAN-originated counterpart to curate-knowledge (which promotes run-originated proposals); both are Tier-3 curated writes and you (the human-in-the-loop) own the truth call. Distilled from graphrag_curated_yaml_roadmap.md + dev log 20260519d (the clumping_index injection).

2026-07-076

markdown-to-pdf

jingtao-lbl/A2MC-elm

Convert a Markdown document (an ana_log, report, or note) to a shareable PDF or Word (.docx) via pandoc. Use when the user asks to "convert/render markdown to PDF/Word/docx", "render this ana_log/report to PDF", "export markdown to a document", "make a PDF of this", "turn this .md into a docx". Prose documents, not slide decks (slides go through Marp). Requires pandoc + a LaTeX engine for PDF, and python-docx for round-trip-safe .docx edits. Self-contained repo copy — depends only on tools the environment provides, not on any user-level skill or ~/.claude workflow.

2026-07-076

name	compare-calibration-rounds
description	Compare A2MC calibration rounds (R1, R2, …) against each other and the validation targets — top-N best-achievable biomass per round, per-target Morris μ* sensitivity, P-pool / cross-regime overlays. Use when the user asks to "compare rounds", "update the multi-round figure", "refresh the R1-RN comparison", "which round is best", "top sensitive parameters per round", or to regenerate any `multiround_` / cross-round figure after new cases complete. Codifies the multiround bundle pipeline + the footguns (screening contamination, case-name patterns, μ ranking, param-set mismatch, partial-ensemble caveats).
modes	{"requires_fates":true,"nutrient_pathway":"any","scope":["analysis"],"summary":"Cross-round comparison of FATES Morris ensembles; requires FATES calibration rounds. Targets from the case targets.yaml."}

Compare Calibration Rounds

When to use

"Compare the rounds", "which round is best", "update/refresh the multi-round figure".
"Top sensitive parameters per round" / cross-round sensitivity.
Regenerating a multiround_* figure after new cases complete (e.g. a round's relaunched/added cases).
P-pool or cross-regime cross-round overlays (same bundle conventions).

The bundle

multiround_top50_sensitivity_20260606/
├── scripts/
│   ├── plot_multiround_top50_overlay.py        # biomass top-50 vs 6 targets (6 panels)
│   ├── plot_multiround_sensitivity_overlay.py  # Morris μ* top params per round (grouped bars)
│   ├── build_Y_from_monthly.py                 # R2..R5 Morris Y matrices from monthly NCs
│   └── build_Y_R1.py                           # R1 (138-param)
├── screening_top50/{R1..RN}/                   # screen_ensemble.py outputs per round
├── sensitivity/{R1..RN}_{leaf,fineroot}/       # morris μ* CSVs + per-round png
├── Y_matrices/                                 # Morris inputs (for sensitivity only)
└── multiround_*.png                            # the comparison figures

Also: tools/compare_rounds.py for quantitative round-vs-round deltas.

Pipeline (to refresh a figure after new cases)

Extract the new cases (all 3 phases if the combined/sensitivity figures are needed): production tools/extract_monthly_variables_FATES.py (TRANS) or tools/extract_and_plot_selected_cases.py (extract, any phase via --phase/--year-start/--year-end, reuses process_case).
Re-screen the round: phases/phase2_screening/screen_ensemble.py --data-dir <round extract dir> --top-n 100 --output-dir screening_top50/<RN> (source that round's config so A2MC_CASE_NAME_PATTERN is right). ⚠️ see footgun #1 — filter out-of-range cases first.
Biomass figure: python scripts/plot_multiround_top50_overlay.py (reads screening_top50/*/).
Sensitivity (only if refreshing μ*): rebuild Y (build_Y_from_monthly.py) → Morris (phases/phase1_exploration/morris_sensitivity_analysis.py) → plot_multiround_sensitivity_overlay.py.

Footguns (the load-bearing part)

#	Footgun	Guard
1	CROSS-round screening must cap the case number. `screen_ensemble.py` matches `PtCNPEn(\d+)<suffix>_TRANS` for any digits, so out-of-Morris-range experiment cases (e.g. H1 clumping #5001/5005/5006) sitting in a shared extract dir get swept in and can rank as "best". (This made the R5 biomass figure briefly show best #5001 instead of #1304, 2026-06-10.)	Pass `screen_ensemble.py --max-case-num <ensemble max>` (e.g. `4890`) for cross-round comparison — it drops out-of-range cases and prints how many. Verify `# Sets:` in `_results.txt` = expected. NOTE: leave `--max-case-num` UNSET for a single round's own whole-ensemble summary graph — those legitimately exceed the Morris size (e.g. R3 + reruns = 4941).
2	Read `_results.txt` Sim_ columns + `screening_result.json` `best_case_num`, NOT `screening_top50_indices.txt`. The indices file mislabels case numbers on partial ensembles and isn't always rewritten (can be stale).	The overlay scripts already do this; don't "trust" the indices file.
3	Morris CSV `rank` column is by `mu`, not `mu_star`.	Re-sort by `mu_star` (Morris importance) for selecting/ranking top params.
4	Sensitivity file prefix is unreliable. Older runs (R1–R4, 6/6) named both organs `morris_leafbiomass_` (the `--output-var` quirk); newer runs (R5, 6/10) correctly use `morris_finerootbiomass_` for fineroot. So a dir can hold either prefix. The directory `{Rn}_{organ}` is the source of truth for the organ; content is correct per dir.	Map organ by directory, never by filename prefix. Glob `morris_PFT{X}_2.csv` (matches both); ensure one file per PFT so `sorted()[-1]` is unambiguous.
5	Param-set mismatch across rounds. R1 used 138 params; R2+ used 162.	Align on shared names; show "no bar"/blank for params absent in a round; note it in the caption.
6	Partial-ensemble rounds. Screening/ranking is valid on a partial ensemble (dead cases get finite high cost), but Morris μ is NOT reliable* until the ensemble is reasonably complete.	Include a partial round in the biomass top-N figure (label it), but exclude it from the sensitivity overlay until a full Y-rebuild + Morris re-run.

Conventions

Filenames embed round + axis-mode + case count: R{N}_{TRANS,combined}_{count}cases_ensemble.png, multiround_top50_biomass_vs_6targets_R1-RN.png, multiround_top_sensitive_params_bars_R1-RN.png (per feedback_plot_filename_convention).
Biomass overlay bands: grey ±20% acceptance band + light-blue ±1 SD observation band (SD often exceeds the mean → floor the lower SD edge at 0). SD values from validation_targets_leafroot.txt.
Whole-ensemble (not top-N) cross-round plots use tools/plot_ensemble_cases.py + regen_ensemble_milestone_plot.sh (different from this bundle's top-50 overlay).
NERSC: scratch (symlink dirs, filtered sets) under $HOME/repo tmp/ only — never /tmp,/scratch.

Cross-references

Worked examples: the originating dev/ana log (procedure), 20260610b (refresh + H1 catch), the originating dev/ana log (findings).
Tools: phases/phase2_screening/screen_ensemble.py, phases/phase1_exploration/morris_sensitivity_analysis.py, tools/{compare_rounds,extract_and_plot_selected_cases,plot_ensemble_cases}.py.

Changelog

2026-06-17: ## Changelog convention adopted (see .claude/skills/README.md). Earlier history: git log + memory/dev_logs/.