Kublai Khan`s Lost Fleet. Koan War (2nd Mongol Invasion of Japan, 1281)

This article will discuss the contribution of Kublai Khan`s lost fleet maritime archaeological site to our understanding of the Mongol invasion force of 1281.  The site is located in Imari Bay, which is in the north west of Kyushu, approximately 48km from Fukuoka.  The site was found and first excavated in the early 1980s by Professor Torao Mozai.  Later in the 1990s a second excavation was carried out, led by marine archaeologist Kenzo Hayashida which continued until the early 2000s.  From these excavations and subsequent analysis of the artefacts together with scholarly research we can learn a lot about the history of the Japan Sea, Mongols, Chinese, Korean and Japanese seafarers.  Respected marine archaeologist and author of, Kublai Khan`s Lost Fleet, James Delgado joined the excavations at Imari Bay in the early 2000s and at the time of writing this essay, has the most up to date information.  (Ed. Just to note though, I wrote this essay about 10 years ago and I had limited knowledge about the Mongol invasions and few resource materials) I was only able to find a few articles, one published in a National Geographic magazine written by Mozai, and a few written by Delgado relating to the excavation of the lost Mongolian fleet and there does not seem to be many journal articles or books written specifically to the marine archaeology of this site.  Jeremy Green`s review of James Delgado’s book also comments on the lack of information relating to this topic and how the book relates to marine archaeology, if it does at all.   

In an analysis of the Japanese site, it is important to discuss the role of the Sung Dynasty and its navy because the wreck site itself has proved to contain many Sung Dynasty vessels which were used by the southern Mongol force in the attempted invasion.  The Mongols themselves never had a navy and the two invasions, first in 1274 and second in 1281, were the first time the Mongols attempted a seaborne attack.  From archaeological evidence of the site, we know that the ships were Chinese made by the Sung Dynasty and Randall Sasaki, a Texas A&M University graduate who joined Hayashida’s team during the excavation, believes the ships found, were at the epitome of Sung naval ship construction in the 13th century.  We also know that the Japanese, although not having a navy were able to use modified fishing boats and turn them into small coastal fireships and also used them as boarding vessels which harassed the invading fleet while anchored in Imari Bay.

Wall built by samurai of Higo (Kumamoto) before the second invasion. This is where the samurai drove off the Mongol attackers. Matsubara, Fukuoka.

Early archaeological excavations were carried out in the 1920s on the ancient defensive wall at Hakata Bay, built between the first and second Mongol invasions, but very little was known about the fate of the lost fleet.  Much later, Mozai conducted some research on the lost fleet but it was not until 1981 that a serious effort was made to undertake a systematic survey and excavation of the Takashima area thought to be where the fleet was lost.  Mozai first visited shrines situated around Hakata Bay in Fukuoka City that are dedicated to the victory over the Mongols and then he visited the Tokyo Museum Imperial Collections where he found a late 13th century scroll that depicts a samurai warrior Takezaki Suenaga fighting against the Mongols. The search had begun.

Samurai Takezaki Suenaga and the 13th century scroll at the Iki no Matsubara site.

Long before archaeological excavations officially started, local fishermen in and around Imari Bay for years had brought up strange artefacts in their nets while fishing, which they suspected had been from the Mongol invasions.  Mozai first went to speak to some of these fishermen and was left speechless after being shown earthenware ceramics and pots that they had collected over the years.  On 14 December 1980 the New York Times reported that the lost fleet of Kublai Khan had been found.  Mozai explained to the New York Times that there are at least 70 wooden hulks that had been found by sonar and that they were sitting in about 1.8m of mud in 24m of water.  He also explained how divers on the site retrieved a sword, stone implements and a bronze statue.   Delgado confirms that among these artefacts were two important pieces that could be traced to the lost Mongol invasion fleet – a bronze Buddha and an inscribed bronze seal that had belonged to a Mongol general, shown below.

Kamakura period, ca. 720 years ago.
(Stamp plate dimensions) Length of each side: 4.5 cm; thickness: 1.5 cm.
Made of bronze. This seal was found on the coast at Kōzaki in 1974. On the back side of the stamp plate is an inscription to the right of the grip, naming a military rank equivalent to a company commander, and another to the left of the grip giving the date of commission as the ninth month of 1277. ‘Phags-pa script was an alphabet which Yuan dynasty founder Kublai Khan ordered the Tibetan Lama Pagba to devise, and was used for official items such as documents and seals. It was used very little after the collapse of the Yuan.

When Kublai Khan became the leader of the Mongol Empire in 1260 the empire stretched from Europe to Asia. The only Asian territories not subjugated at that time were Korea and Southern China to which the Mongols quickly moved.  The Sung Empire in Southern China was the leader of seafaring and had the most powerful navy in the world.  Khan and the Mongol forces campaigned in Korea and China taking control of the Sung navy with Sung defectors and eventually defeated the Sung Emperor in 1279.  In July of the same year a Chinese refugee arrived in Japan warning the Japanese of the fall of the Sung dynasty and that the new Mongol rulers had established the Yuan Dynasty and once again had its sights on Japan.  Further expansion across the sea to Japan was now made possible with the Khan’s acquisition of the Sung navy and manpower.  Khan established a new government office called the Ministry for Conquering Japan and ordered the Southern Chinese of the Yangtze River to build 900 ships.   Khan decided on the first month of 1281 for the next invasion of Japan.

1281 invasion routes according to Hakkutsu sareta retto 2012, Agency for Cultural Affairs.

The sea lanes between Korea and Japan that had been used for centuries for trade and piracy now became sea lanes for invasion.  There were two fleets consisting of an eastern division from Korea and a southern division from China. The Eastern division consisted of 10,000 Koreans and 30, 000 Mongols, 900 ships and 17,000 crew which left from Happo in Goryeo (now Masan near Busan) on May 22nd, commanded by Mongol General Hong Dagu and Korean General Kim Bang-gyong.

The southern Chinese division was the main invasion force, leaving from Qingyuan (modern Ningbo) at the beginning of July. The southern division consisted of 100,000 Sung soldiers and 2,600 warships commanded by Chinese Fan Wenhu and overall invasion commander Mongol General Arakhan. But from the outset the invasion was doomed to fail. Lets see why.

The Eastern division was much more prepared and eager to get going than the main invasion force in China. It seems that the Chinese were still struggling to assemble the number of ships and morale was running low due to many years of war. There was also underlying competition between the Mongol commanders which plays out during the invasion.

The Korean fleet attacked Tsushima on 9th June and Iki on the 14th. Now according to different sources, both fleets were to meet at either Iki or Hirado and then make make a combined attack on Hakata. I will assume the meeting point was Iki for this essay. The southern fleet was expected at Iki by the 2nd July. However, ignoring orders to wait, the Korean force left Iki and headed to Hakata. However, for some unknown reason, on the way, a detachment of 300 ships, which was a third of the Korean force, was sent off to attack the Nagato region, now Yamaguchi and made their attacks on 25th June. Little is known about these attacks, and to be honest, we are left scratching our heads why this decision was made. It reduced manpower even more, for the Hakata attacks. The only way the invasion had any chance of success was with a full force, and for now, the Mongols attacked Hakata with only about 26,000 troops instead of a possible full force of 140,000 men. Of course we can look in hindsight but the Mongol commanders also knew that the Japanese had built a wall and that the samurai were waiting for them. Was it arrogance, stupidity or was general Dong Dagu aiming for fame over his compatriot? We may never know.

In the meantime the main force arrived in Hakata, on the western side of Hakata Bay, pictured above, and the invasion began on 23rd June. Now, the Japanese had been ready for an invasion since about spring of 1280. Samurai had erected a wall, at least 1.8m high, and earth ramparts were built behind the wall that the archers could stand on which enables them to fire at the attackers without exposing themselves to enemy fire.  The Japanese forced the invaders into close quarters fighting, between the wall and the ocean which was not how the Mongols usually fought. Mongol cavalry was the best in the world but needed space to be able to harass and charge at the enemy.  The stern Japanese defence resulted in the Mongols not being able to make a landing, even after 50 days of fighting, so instead retreated to Shikanoshima and Nokonoshima, two small islands in Hakata Bay.  The Japanese defenders, then turned attackers and chased the Mongols.  The Japanese used small coastal ships to attack, harass and board the enemy ships.  Samurai warriors who were much more accustomed to close quarters fighting managed to climb aboard the enemy ships and do a great deal of damage killing the ships crews and soldiers.  After nearly two months of fighting the attackers, tired and with morale very low, the Mongol commander decided to give up on Hakata, retreat back to Iki, and go look for the main Chinese force.

Samurai attacking the Yuan ships by small boats.

The defeated Eastern force finally met up with the southern Chinese fleet at the beginning of July. Some say they met up at Iki as ordered, but other accounts say they met around Takashima, which makes sense. Also, at this point it is unclear whether the Mongol combined forces planned to land and attack in the Imari bay area as Delgado suggests in his article or whether they were on their way back to Hakata Bay as Ishii suggests.  I will assume that news from the Eastern army of the fortified wall and heavy resistance experienced just weeks before at Hakata Bay was enough of a sign for them to find an alternate landing place.  Therefore, the Mongol fleet arrived at Imari Bay near Takashima Island and on the 1st July where they prepared for their final combined invasion.

The southern fleet was primarily made up of Chinese junks that were used by the Sung Dynasty for trade throughout South Eastern Asia.  Wooden artefacts believed to be from the sunken ships found at the site have enabled Japanese archaeologists to make computer simulated ships that may have been used in the invasion force. Image below from Kosuwa.

Computer generated image of the Chinese invasion ship.

The majority of Chinese and Japanese coastal vessels used in this conflict were primarily suited to short coastal voyages in and around the Japan Sea and the Japanese Archipelago.  There were a few bigger ships that resemble ocean going vessels, but the majority were smaller vessels.  It is thought from the artefacts gathered at the Takashima site that the ships were shallow hulled, 2 mast vessels. They were made from Chinese oak and pine identifying their place of manufacture and were designed to carry a lot of soldiers and horses and were equipped with watertight bulkheads, a rudder, an anchor, a compass and explosive projectiles chests.  The discovery of an intact anchor at the site gives a greater understanding of ocean and sea conditions this fleet was expected to face.  One would assume a heavier sea and swell would require the need for a larger and sturdier anchor.  The anchor found at the bottom of the harbour was small and light in weight and inferior to the known Chinese anchors of the time.  Delgado notes that the usual single large granite weight usually found on Sung ship anchors were not used and that the anchor as shown above was crudely made with 2 smaller granite stones and a short stock which did not have the strength to hold the ships at anchor when the typhoon passed through and the sea smashed them against each other.  This also raises a couple of questions.  Were the Chinese short of materials to build a sturdy fleet and were the ships built in a hurry that neglected seaworthiness.   

One of the most important finds on the site were those of the ceramic bombs.  I would like to believe this invasion force was the first ocean going invasion that used gunpower and projectile weapons in human history. 

Ceramic hand thrown bomb as used by the attacking Mongols

If it is, it is definitely a turning point for human warfare in the relation to seaborne assaults.  The 13th century scrolls that I mentioned earlier depict Mongol warriors throwing these hand bombs at Japanese defenders, exploding in mid air.  X-rays on these artefacts reveal some of these weapons to be loaded with gunpowder and other loaded with gunpowder and bits of metal shrapnel.

Wooden anchor found at the site.

On the eve of their final invasion and after weeks of constant attacks by the Japanese the Mongol generals decided to chain their ships together to form a huge floating fortress to thwart the Japanese coastal fire vessels and samurai hit and run attacks which although small on the scale compared to the Mongol fleet were enough to keep them busy while they prepared for the invasion.  Wooden ship fragments which have survived over 700 years below the surface of the sea still show deep scaring and scorching which prove the fleet had suffered fire damage either by the Japanese attacks, or that the ships caught fire during the course of the typhoon.  Fires may have also been ignited by blacksmith forges that the Mongols used on the ships to make and repair metal objects such as horse shoes, armour and weapons.  Many bricks were found by Mozai in his first excavation of the site which he believes may have been used to make the portable forges.

The gods answered the prayers of the Shinto priests who had been performing elaborate ceremonies on behalf of the defending Japanese.  A great “Divine Wind” (typhoon) hit Kyushu and the Imari Bay area.  As the typhoon began to intensify throughout the day the fleet unchained themselves from each other and made for the open sea.  Mongol lack of knowledge with the sea may have contributed to what appears to be a slow response to the coming storm.  The typhoon hit quickly with high winds, driving rain and heavy seas.  But due to the mass of ships floating in the bay, most were unable to escape and they began to smash into each other and those that did not hit other ships were driven against the coastline.  The storm was strong enough to destroy the Mongol fleet causing the loss of over 4000 ships and an estimated 100,000 dead soldiers.

The New Yorks Times ran a second story on 30 August 1981, a year after the first story which broke the news of Mozai finding the lost fleet.  More artefacts brought up from the bottom confirmed what Mozai had hoped and this story published by the New York Times shows the progress of the excavation and how successful it is.  The expedition in less than a year uncovering enormous numbers of artifacts and proving that the discovery is the most important ever found in Japanese waters.  Mozai explains that a team of 20 divers have brought up enough pottery and weaponry to fill 10 large suitcases.  Also found were 145 fragments of what are believed to be vases, some of which were used to mix and store gunpowder.

For 700 years the sea and a whole lot of mud had kept a secret.  Hundreds of ships, weapons, pots, bowls, vases, and human remains lay in 1.8m of ocean floor silt.  Archaeologists dug through the mud of Imari Bay to find the seabed scattered with the remnants of the Mongol lost fleet.  It was and is the biggest invasion force ever assembled before the Normandy invasion during World War 2.  By the 13th century the Sung Dynasty had amassed a huge and powerful navy protecting the trade lanes that stretched from the Indian Ocean, the Yellow Sea and all through Asia.  The Mongol conquest of the Sung gave them the the means to attempt another invasion on their old foe and the Japanese archipelago.  However, the marine archaeological site of Kublai`s lost fleet in Imari Bay has shown that even the best invasion force can be defeated by nature and lost to history.  


Delgado, J.  2003. Relics of the Kamikaze.  Archaeology.  56 (1),  Archaeological Institute of America.

Delgado, J. 2003. Shooting down the Kamikaze myth. Naval History, 17(3), pp. 36-41.

Delgado, J.  2011. Kublai Khan VS Kamikaze, Military History, 28 (2), Academic Research Library

Delgado, J. 2008. Khubilai Khan`s Lost Fleet. University of California Press.

Green, J.  2010.  Review of the book Kublai Khan’s Lost Fleet by James Delgado, International Journal of Nautical Archaeology, 39 (1)

Ishii, S.  1995.  The Decline of the Kamakura Bakufu.  In Jansen, M (ed.), Warrior Rule in Japan. pp. 44-76. New York, Cambridge University Press.

Mozai, T.  1982.  The Lost Fleet of Kublai Khan, National Geographic, 162, pp. 635-649

Turnbull, Stephen. 2010. The Mongol Invasions of Japan 1274 and 1281. Osprey.

Yamada, Nakaba. 2012. Ghenko.  The Mongol Invasion of Japan 1274 – 81.  Leonaur.

Own photos and artefact photos courtesy of the Japanese archaeology association.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s