The Doublet Is Real.
“Same Spot” Was False Precision.

What Bath's Law Says

When a big earthquake breaks on a fault, the stress it releases triggers smaller quakes on nearby patches. A statistical regularity called Bath's law, known since 1965, predicts that the largest aftershock is typically about 1.2 magnitude units smaller than the mainshock. For a M6.0, that means an expected largest aftershock of roughly M4.8. The deviation around that figure has a standard deviation of about 0.4, so M4.4 to M5.2 is the normal band.

The Kermadec M5.9 sits 2.75 standard deviations above the Bath's-law expectation. Under the null hypothesis that this pair is a standard mainshock-aftershock sequence, the probability of seeing a gap this small is p = 0.006.

The Networks Agree on Size

The USGS, EMSC, GFZ, and ISC catalogs were cross-checked to confirm both events exist and agree on magnitude. They do — to within a few hundredths of a unit.

That agreement on size is exactly what makes the doublet finding solid — Bath's law is a statement about magnitudes, and every network sees the same two near-equal quakes.

But Not on Where

How close together were they? “Same spot” is the tempting answer — an early solution puts the two quakes about 12 km apart. But that precision is an illusion. This is an offshore sequence, hundreds of kilometres from the nearest land seismometer, and locating offshore quakes is genuinely hard — the arrival-time geometry is one-sided, so every agency's epicenter carries an uncertainty of tens of kilometres.

When you actually ask each catalog where the M6.0 was, they scatter across ~70 km of longitude. USGS's own solution moved 25.8 km between its first automatic pick and its reviewed determination. And the separation between the two quakes — the thing “same spot” is really about — runs anywhere from 14 to 26 km depending on which solution you trust:

None of this touches the doublet result. The pair is still a genuine doublet — two near-equal ruptures, 45 hours apart, in the same offshore patch of the arc. What the data can't support is a claim about their exact shared point.

Doublet, Not Aftershock

An earthquake doublet is a pair of similar-magnitude events on adjacent segments of the same fault system, triggered by the stress transfer of the first onto the second patch. Doublets are rare. The most cited example is the 2009 Samoa M8.1 doublet, where a pair of M8.0-range events ruptured within minutes on different fault systems. The 1933 Sanriku and 1971 Kuril Islands sequences also qualify.

The Kermadec subduction zone, where the Pacific plate dives under the Australian plate, is one of the fastest-converging plate boundaries on Earth. Multiple adjacent patches of the interface can sit near their rupture threshold at the same time. When one slips, the stress jumps to its neighbor.

Figure: Magnitude Through Time

What the Result Does Not Say

The Bath's-law rejection is a point comparison against the literature value of 1.2. It rests on N = 5 aftershocks within 30 days, which is a small sample. To check whether the post-M6.0 aftershock population is statistically anomalous as a whole, we ran two distribution-level tests against the historical Kermadec M5 and above catalog for 2011 through April 2026 (N = 142).

Neither test rejects the null. The distribution of aftershock magnitudes is consistent with the baseline Kermadec M5+ catalog. Our claim is narrowly that the largest aftershock is far above the Bath's-law expectation, not that the aftershock sequence as a whole is anomalous. The doublet identification sits on the point comparison, flagged for verification once 60 to 90 more days of catalog data accumulate.

The Sensitivity Table

To check whether the result depends on arbitrary choices, the Bath's-law test was re-run across eight variants of the analysis window: three distance radii (100, 200, 300 km), three magnitude-completeness thresholds (Mc 2.5, 3.0, 3.5), and two alternative time windows (14 days, 60 days). All eight variants return identical results: N = 5, Δm = 0.10, p = 0.006. The two events dominate any reasonable analysis window.

What We Still Cannot Do

We cannot say whether the M6.0 and M5.9 ruptured the same fault patch or adjacent ones — that needs a focal-mechanism solution and slip inversion, beyond catalog-level analysis, and the ~tens-of-km location scatter means even the pair's separation isn't pinned down. We cannot forecast whether more doublet-range events are imminent. We can say that the pair is statistically inconsistent with a standard mainshock-aftershock sequence, and that the independent networks agree on the magnitudes to within 0.03 units.

Reproducibility

All code, data extracts, and the full statistical pipeline live in the kermadec-doublet-m6-0-m5-9-same-epicente workspace. Every number on this page comes from data/results.json. Seed: 42. Pipeline: USGS FDSN + EMSC + GFZ + ISC catalog pulls, earthquake deduplication, Bath's law point test, Gutenberg-Richter b-value via Aki MLE, Omori-Utsu aftershock decay fit, Welch t and Mann-Whitney U on the distribution, Bonferroni correction across the test family.