P4: survey hygiene (S-CD.1-3, SCI-08): Obermann key collision, denominators, reporting rates #42

Package P4 of the iteration-3 plan: the literature survey's denominators and citations. Owner: agent; the author verifies rows in #48. Size: a day.

Closes or advances: S-CD.1, S-CD.2, S-CD.3, SCI-08. Check existing issue #30 (dangling keys @lobkis01, @poupinet84) against the current build; the iteration-2 inventory found every key resolving.

• Key collision (S-CD.1). paper/build_survey.py make_key() yields Obermann2013 for two different papers (JGR 10.1002/2013JB010399, Piton de la Fournaise; GJI 10.1093/gji/ggt043, numerical media) and skips disambiguation when the key exists in references.bib. Compare the row DOI with the bib entry's DOI even for reused keys; disambiguate to Obermann2013b (the Crossref cache holds the JGR record); regenerate survey.bib and appendix_table.tex; the reference list must then contain both papers.
• Counts. Write "103 rows for 103 publications" and reconcile qmd 161, 787 and the table caption with that count and with the 7 rows that self-declare no dv/v measurement and the 11 earthquake-coda-only rows (state the rule or exclude them).
• Verified reporting rates (S-CD.3, SCI-08). Report per-field rates for the full-text dv/v rows with a coding rule that separates an error estimate from a QC threshold from n/r (from the CSV: band 69/82, window 59/82, method 74/82, uncertainty 72/82 non-n/r before the split); say that 21 rows are abstract-derived and excluded from the rate; use one noun (rows, publications or studies) throughout.
• Search rules. One sentence each for search dates, queries, inclusion rules and deduplication in Appendix C.
• Yuan regime (S-CD.2). After Author input: confirm Yuan et al. (2021) Table B3 estimator ranking (blocks P1, P4) #47 confirms Table B3, state the regime (|dv/v| <= 0.5 percent, 0-30 percent added noise; WTS and MWCS named most noise-robust) where Results 346-348 cites the ranking.

Acceptance: build_survey.py has a test for the DOI-aware key; the appendix table, the prose counts and the caption agree; both Obermann 2013 papers appear in the rendered reference list.

Working agreement (agents and humans). One PR per issue, branch iter3/<package> off docs/sign-convention-manuscript; name the ledger IDs in the PR title; request Copilot review before approval. Do not edit review/REVISION_PLAN_ITER3.md except to tick boxes and append a dated log line naming this issue. Do not hand-edit statuses in reviews/codameter-gji.review.json; the reviewer reconciles them in P7. Tests and pre-commit hooks must pass. Before requesting review, rebuild the PDF (cd paper && ../.pixi/envs/dev/bin/python build.py --no-survey) and check its text for ??, ,%, << and [Funding. Evidence for every finding: reviews/iter2/block_S-*.md; full report reviews/codameter-gji.iter2.report.md; plan review/REVISION_PLAN_ITER3.md.