Is Japan shaking more?

The timeline — every quake, five years

Each dot is one earthquake. Height = magnitude. Colour = depth (red = shallow, blue = deep). Hover any dot for details. Filter by year or zoom to M6+ only.

The 2011 Tōhoku M9.1 was the last event M8.5+ in this region — nothing in the last five years comes close. The two headline events are the 2024 Noto Peninsula M7.5 and the Dec 2025 Aomori M7.6.

Frequency — are quakes getting more common?

Monthly count of M4.5+ events, with a 12-month rolling average. Toggle to the M6+ view for the strong-quake cadence.

The +9% recent-vs-earlier mean is within normal year-to-year fluctuation. Japan's M4.5+ rate has bounced between ~500 and ~600 per year with no sustained climb.

The strong ones — M6+ and M7+ vs a 20-year baseline

This is the decisive chart. The grey bars: M6+ quakes per year 2002–2020 (19 full years, pre-window baseline). The orange bars: the four full years in our window. The giant red bar is 2011 — Tōhoku and its aftershocks — an extreme outlier.

Interpretation: on strong-quake frequency, the last five years are statistically indistinguishable from the 20-year average — if anything a touch quieter than the long-run mean once 2011's aftershock storm is set aside.

The biggest events

Top 8 by magnitude in the five-year window. Two stand out — both M7.5+, both recent.

The verdict

Question: are we seeing more frequent and stronger earthquakes in Japan?

Method & caveats

How this was built, and where the numbers can mislead.

• Source: USGS FDSN earthquake catalogue, queried 10 Aug 2026. Region box: lat 24–46°N, lon 125–148°E (Japan proper + surrounding arcs). Window: 12 Aug 2021 → 8 Aug 2026, all events M4.5+ (2,842). Baseline: 2001–2021 M6+ (259 events).
• Completeness: M4.5+ is fully recorded in this region for the whole window. Going further back, detection thresholds were higher, which is why the baseline comparison is done on M6+ only — M6+ has been complete since the 1990s.
• Magnitudes: USGS mixes moment magnitude (Mw) and other types for large events; JMA's own magnitudes occasionally differ by 0.1–0.3 (e.g. JMA rated Aomori M7.6, USGS M7.6). The relative trend conclusions are robust to these differences.
• 2011 distortion: the 84 M6+ events of 2011 are mostly Tōhoku aftershocks — including them in a mean misleads; the comparison above uses both means, with 2011 excluded as the fair baseline.
• Felt vs recorded: JMA reports ~2,000 felt earthquakes per year in Japan; this chart counts recorded M4.5+ events, which is the right metric for trend analysis but a much smaller number than what people feel.
• Short window: five years is a snapshot. Regional stress changes (e.g. the 2024 Noto swarm) can raise activity locally for months; a cluster is not a trend.

Sources