Skip to content
Science Journaling Club

INTERACTIVE MODELFIELD NOTEGEOPHYSICS

The Re-Occupied Position

A live model accompanying “The Iron Ball at the Center of the Earth Keeps Changing Its Skin”

Read the full field noteOur calculation

Study
Field note
Review
Peer-edited by the club review board
Article
13 sectionsabout 23 min
Files

The whole method in this paper depends on one trick: the inner core's own drifting rotation eventually carries it back through a position it already occupied years before, and when that happens a seismic wave gets to re-sample rock it already sampled once. Drag through the schematic rotation history below until you find one of those moments, then decide whether the shallow boundary changed in the meantime, and watch two seismic arrays disagree about whether it did.

Model 1. Find a repeated position

This curve is the schematic shape of the rotation history reported in the 2024 backtracking paper — five years of faster-than-mantle "super-rotation" (2003–2008), then roughly three times as many years of slower "sub-rotation" moving back the other way (2008–2023). The vertical axis has no real units attached to it; the paper reports timing and a rate ratio, not a single plottable number. Drag the year slider through the descending part of the curve and it will find the earlier year, during the ascending part, where the inner core sat in the same relative position.

phase: before super-rotationfirst occupied this position: –years apart: –
Drag past 2008 to start finding repeats. Every year in the sub-rotation phase (shaded) matches some earlier year in the super-rotation phase — that match is a usable pair.

Model 2. Two ears, one boundary

Once you have a repeated position, the paper's actual question is simple: do the two recordings, years apart, still match? The Eielson array (ILAR) samples a ray path that bottoms out deep in the inner core. The Yellowknife array (YKA) samples a shallower, boundary-grazing path. Toggle whether the shallow boundary changed between the two visits and watch what each station would see.

Has the shallow boundary changed since the first visit?
ILAR (deep path): matchYKA (shallow path): match
With no change near the boundary, a ray that bottoms deep and a ray that grazes the rim both re-sample exactly the same undisturbed material. Both match. This is what the vast majority of repeated positions actually look like.

Model 3. How big, how thin

Two real numbers, neither stated outright in the paper: how large the inner core is, and how thin a structurally "soft" layer at its edge can be and still be a real, published result. Drag the thickness slider across the range an unrelated 2017 study actually reported for a mushy patch elsewhere on the same boundary.

Inner-core radius (PREM)
1,221.5 km
Inner-core volume
7.63 × 109 km³
As a share of the Moon's volume
≈ 35%
This layer's share of the radius
0.491%

Tian & Wen (2017) reported a localized mushy zone 4–8 km thick beneath the southwest Okhotsk Sea — a different patch of the same boundary, found with a different seismic method. It is not the thickness of whatever changed at YKA between 2004 and 2008, which the 2025 paper does not quantify in kilometres. Drag inside and outside that 4–8 km range to see how little of the radius is at stake either way.

What the sliders are trying to show

Two things tend to land harder here than in prose. First, the rotation trick in Model 1 is the entire reason this measurement is possible at all — without a slow reversal handing back old positions, there would be no way to separate "the ray moved" from "the rock changed." Second, Model 2's asymmetry is the paper's whole argument compressed into two buttons: the same underlying event, recorded twice, at two stations that sample different depths, is the only way to say where in the inner core something happened rather than just that something did.