Nature Geoscience Supplementary Information (ngeo-supplementary)

When to trigger

• The Article is over 3,000 words and detail must move out of the main text
• A reader cannot follow the main text without consulting the Supplementary Information
• You have extended maps, full model ensembles, derivations, or additional proxies with no home
• You are unsure what is allowed in SI versus what must stay in the main text or online Methods
• Referees of a prior version asked for material that would bloat the main text

The stand-alone rule (non-negotiable)

The main text must be fully understandable on its own. The Supplementary Information supports the Article; it must never be load-bearing for the headline advance. A reader who never opens the SI should still grasp the result, the core evidence for it, its uncertainty, and why it matters.

• The headline advance, its decisive evidence, and the headline uncertainty live in the main text.
• The online Methods holds the reproducible protocol (instruments, model setup, statistics).
• The SI holds the exhaustive backup a skeptical referee or replicator wants: extended datasets, full ensembles, sensitivity tests, additional figures and tables.

If removing the SI breaks the argument, the partition is wrong — pull the essential piece back into the main text (and trim elsewhere via ngeo-length-management).

Methods vs. SI (a common confusion)

Nature Geoscience has both an online Methods section and Supplementary Information; they are not the same:

SI organization

• Mirror the Article's logic: order SI sections to match the order they are cited in the main text and Methods.
• Give SI sections clear headings and self-contained captions; label figures/tables Fig. Sn / Table Sn / Supplementary Note n.
• Every SI item is cited from the main text or Methods at the point it is relevant; nothing goes uncited.
• Keep notation, units, and symbol definitions consistent with the main text and Methods.
• Keep the SI honest: it is peer-reviewed alongside the Article, not a place to bury weak results, unvalidated model runs, or claims the referees did not vet.

Checklist

• The main text is understandable without ever opening the SI
• No headline-advance evidence is only in the SI
• Reproducible protocol is in online Methods, not scattered in SI
• Every SI item is cited from the main text or Methods at the relevant point
• SI sections ordered to match their citation order
• SI figures/tables labeled Fig. Sn / Table Sn distinctly
• Extended data, full ensembles, and derivations live in SI
• Notation and units consistent across main text, Methods, and SI

Anti-patterns

• A main text that silently depends on the SI to be understood
• Burying the decisive evidence or headline uncertainty in the SI
• Confusing SI with online Methods (protocol scattered instead of consolidated)
• An SI that is a disorganized dump rather than a structured appendix
• SI items referenced out of order or never cited
• Notation/unit drift between main text, Methods, and SI
• Using SI to smuggle in unvalidated model runs or claims reviewers did not see

Output format

【Main text stands alone】yes / fix (pull back: ...)
【Methods vs. SI split correct】yes / fix
【SI sections】S1: ... / S2: ... (cited from main text/Methods?)
【Headline evidence only in SI?】none / move back: ...
【SI labels】Fig. Sn / Table Sn?  yes / fix
【Notation consistent across all three】yes / fix
【Next】ngeo-writing-style or ngeo-length-management

SI scope, labeling, and the Methods/SI boundary evolve — verify on the official Nature Geoscience author pages (Checked: 2026-07-16).