SEATTLE - Evidence has mounted for nearly 20 years that a great earthquake ripped the seafloor off the Washington coast in 1700, long before there were any written records in the region. Now, a newly authenticated record of a fatal shipwreck in Japan has added an intriguing clue.
Written records collected from villages along a 500-mile stretch of the main Japanese island of Honshu show the coast was hit by a series of waves, collectively called a tsunami, on Jan. 28, 1700. Because no Japanese earthquake warned of the waves, it is likely they came from somewhere else around the Pacific Rim, said Brian Atwater, an affiliate professor of Earth and space sciences at the University of Washington and a U.S. Geological Survey geologist.
In the village of Kuwagasaki (now part of the town of Miyako) 300 miles northeast of Tokyo, the tsunami is believed to have crested at about 10 feet, destroying 13 houses and starting a fire that consumed additional houses. Records from five other towns lend more evidence for a tsunami generated by a magnitude 9 earthquake off the Washington coast on Jan. 26, 1700.
The shipwreck story is different from other accounts, said Atwater, who will present evidence of the incident Tuesday at the Geological Society of America annual meeting in Seattle.
"This is the only account that is nautical, and it is the only one in which the tsunami contributed to deaths," he said.
The ship carried 470 bales of rice, nearly 30 tons in all, bound for Edo (now Tokyo) from Nakamura-han, a feudal domain. The ship was loaded on Tuesday, Jan. 26, and had sailed about 100 miles down the coast by pre-dawn hours on Thurs-day. It was to enter a river at the town of Nakami-nato, where the rice would be transferred to skiffs for the rest of the journey to Edo.
However, a series of then-puzzling waves sloshing into and out of the river on Thursday created conditions too treacherous for the ship to negotiate rocks just beyond the river's mouth, so the crew kept the vessel anchored just offshore. Atwater likened the situation to the dangerous conditions that often accompany strong ebb tides across sandbars at the entrances to rivers, creeks and harbors.
"This was going on during the early morning hours," he said. "The boat stayed offshore all day. It seems the tsunami lasted 18 hours at least, and that's another hint that the earthquake that caused it was very big."
By the time the waves subsided, a large storm was brewing. The tsunami had kept the ship from safe harbor, leaving it at the mercy of the storm. In the high wind and rough sea, the vessel broke loose from its anchor lines. The crew lightened the load by throwing half the rice overboard, but the storm drove the ship into coastal rocks, two crew members died and all the rice was lost.
An account of the disaster was published in a 1943 book about Japanese shipwrecks, but the source of the document wasn't listed and the account was later met with some skepticism. However, in 2002 Kenji Satake, a Japanese geoscientist, found that the story had been collected as part of a Nakaminato municipal history, and he traced that account to a local family that had kept records of 131 shipwrecks between 1670 and 1832. The account matches well with the other records that supply evidence of the tsunami, Atwater said.
To judge the 1700 earthquake's size, its estimated magnitude can be compared with those of the 20th century's largest quakes - the 1952 Kamchatka earthquake at magnitude 9.0, the 1960 Chilean earthquake at 9.5 and the 1964 Alaskan earthquake at 9.2. The size of the 1700 tsunami in Japan implies that quake was in the range of 8.7 to 9.2, Atwater said, most probably about 9.0.
That quake is believed to have ruptured more than 600 miles of the boundary between the Juan de Fuca and North American tectonic plates, which meet off the Pacific Northwest coast in what is called the Cascadia Subduction Zone. Atwater has previously reported finding a variety of signs of such a large earthquake along the Washington coast, and has gradually pinpointed the event to 1700. The realization that such a large quake could happen again, he noted, has altered building codes and emergency plans in the Pacific Northwest.
At a magnitude of 9.0, the energy generated by the 1700 quake would have exceeded the total amount of energy currently consumed in the U.S. in a month, Atwater said.
"If you were to let a hurricane like Isabel, earlier this year, run for 70 days, the energy released in the winds would be equivalent to a magnitude 9 earthquake," he said.