"It was the phenomenon just as I had envisioned it," says the 41-year-old geologist, who has now become the Japanese Cassandra.
Dr. Shishikura's studies of ancient earth layers persuaded him that every 450 to 800 years, colliding plates in the Pacific triggered waves that devastated areas around the modern city of Sendai, in Miyagi Prefecture, as well as in Fukushima Prefecture.
One early tsunami was known to historians. Caused by the 869 Jogan quake, its waves, according to one chronicle, killed 1,000 people. Dr. Shishikura had found strong evidence of a later tsunami in the same region, which probably took place between 1300 and 1600.
"We cannot deny the possibility that [such a tsunami] will occur again in the near future," he and colleagues wrote in August 2010. That article appeared in a journal published by the Active Fault and Earthquake Research Center in Tsukuba, the government-funded institute where Dr. Shishikura works.
He was beginning to spread the word. Plans were under way at his center to hand out maps so people would understand which areas were at risk. Dr. Shishikura had an appointment on March 23 to explain his research to officials in Fukushima.
Dr. Shishikura's boss at the center, Yukinobu Okamura, had even mentioned the results at a 2009 meeting of an official committee discussing the safety of nuclear-power plants. Dr. Okamura says the idea of beefing up tsunami preparedness didn't go anywhere.
At Dr. Shishikura's eighth-floor office, bookshelves and televisions crashed to the floor during the quake on March 11. He has found temporary office quarters one story below, where he discussed his unheeded warning. "It's unfortunate that it wasn't in time," he said. But he also felt vindicated after past slights, remembering the local official who didn't want to help him dig holes in the earth for research and who called the endeavor a "nuisance."
His work is part of a young field called paleoseismology. Kerry Sieh, a pioneer in the specialty, says that the few dozen people who do this kind of work are usually doomed to be ignored. Humans are made to trust what they have seen themselves, or what someone they know has seen. They aren't designed "to deal with these once-in-500-year events," says Dr. Sieh, formerly of the California Institute of Technology and now head of the Earth Observatory of Singapore.
From his youth, Dr. Shishikura liked to collect fossils in the hills outside Tokyo. He says he realized in high school how geology could answer questions about the past.
His method is fairly simple. Miyagi Prefecture has rich soil, but sandwiched in it are layers of sand and pebbles that Dr. Shishikura says must have been carried from the shore by tsunamis. Looking at the layers allowed his group to estimate the rough dates of waves that struck as far back as 3,500 years ago.
Many lives could have been saved, at relatively little cost, by spreading awareness of the danger. People in Miyagi and Fukushima prefectures were used to strong quakes, but the location and magnitude of these seismic events didn't generate tsunamis. Further north on the eastern coast, tsunamis were well-known from quakes in 1896 and 1933. Those were of yet another, weaker variety that affected mainly low-lying areas along the coast.
During the magnitude 9.0 quake on March 11, some people well inland, thinking themselves safe, took time to change clothes or to make phone calls. Others watched the disaster unfold instead of running to high ground. They proved what Dr. Shishikura's group wrote last year about local tsunamis: "It appears to be almost completely unknown among the general public that in the past great tsunamis have inundated areas as far as 3-4 kilometers inland as the result of earthquakes exceeding magnitude 8."
Now, Dr. Shishikura's team is looking at the Nankai trough to the south, which could trigger tsunamis hitting the island of Shikoku and the Kii Peninsula. Dr. Shishikura says large tsunamis appear to hit there every 400 to 600 years, with the most recent in 1707.
Those rough calculations suggest the danger is at least a century away. Still, Dr. Shishikura says, "we had better be on the lookout."