August 15, 1281

Mongol invasions of Japan

Typhoon devastates the second Yuan invasion fleet off Kyūshū

Underwater archaeologists have identified two invasion shipwrecks beneath the seabed off Takashima; both were built in southern China.

Mid-August 1281

A powerful tropical cyclone struck Kyūshū and the Yuan–Goryeo invasion fleet on or about 15–16 August 1281, sinking, grounding and scattering many ships.

The converted Western date is not uniform: NOAA gives 15–16 August, while other modern works give 13–14 August, 14 August or 23 August, reflecting different conversions and chronologies.

Much of the combined fleet was gathered around Takashima and Imari Bay, and both period records and underwater archaeology associate the destruction with that location.

The storm was not the whole story

The invasion did not fail through weather alone. Before the storm, Japanese fortifications, coastal resistance and small-boat attacks had prevented a secure landing and a sustained bridgehead.

The two divisions had converged around Hirado and Takashima after weeks of unsuccessful operations against northern Kyūshū, leaving a large, crowded fleet exposed near shore when the cyclone arrived.

The Southern Route fleet did not arrive first; the Eastern Route force from Korea had sailed and attacked earlier.

Numbers that will not verify

The traditional totals of more than 140,000 soldiers and 4,000 ships come from textual traditions. Thomas Conlan argues the manpower figure may be inflated by a factor of ten or more.

Archaeology confirms major destruction but does not verify claims that 4,000 ships or 80 per cent of personnel were lost. One modern simulation puts the likely maximum wave height at roughly four to six metres.

Kublai did not immediately abandon Japan: he renewed preparations for another invasion before diverting resources toward Champa and Đại Việt.

Sources

Researched 24 Aug 2026 8 sources not yet audited Date disputed

How this was checked

Researched from the web into a fact sheet and rewritten from that sheet. This article has not yet had an independent model audit. How the pipeline works →

Date disputed. Sources disagree about when this happened.

What the sources leave uncertain

The storm’s converted Western date is not uniform. NOAA gives 15–16 August; other modern works give 13–14 August, 14 August or 23 August, reflecting different conversions and chronologies. The traditional event date should therefore not be treated as hour-level precision.
The exact place where the storm caught every fleet element is uncertain. Archaeology strongly supports destruction around Takashima in Imari Bay, while some narratives place ships at Hakata Bay, in the Tsushima Strait or already withdrawing toward Korea.
The traditional totals of roughly 140,000 personnel and 4,400 ships derive from textual traditions. Thomas Conlan argues that the manpower figure may be inflated by a factor of ten or more.
Ship and casualty percentages cannot be reconstructed securely. Archaeology confirms major destruction but does not verify claims that 4,000 ships or 80 percent of all personnel were lost.