CHAPTER TWENTY-ONE: EARTHQUAKE NOTES

For all their indiscriminate horror, earthquakes seem to have had little effect on the flow of history ‹ Claudio Vita-Finzi

Santa Cruz in 1989

(largely based on San Francisco Chronicle article)

The Pacific Garden Mall, in downtown Santa Cruz, was a key element in the economy of the city. Extending for six blocks near the city's only major tourist attraction, the Santa Cruz Boardwalk, it represented out-of-town dollars for a city that had never been really affluent. The mall generated more than $100 million a year in retail sales, 20% of the city's total sales volume.

City administrators fostered the economic survival of the mall at the expense of earthquake safety, and it was left virtually defenseless against a large earthquake. Conscious decisions were made to put off the capital investment that would have strengthened the old masonry buildings. As a result, the devastation in the earthquake was immense. At least a third of the mall's 75 commercial buildings were damaged beyond repair; almost 2000 people were put out of work; low-income housing was lost; and more than 200 businesses were displaced. And the three people who were killed in Santa Cruz died because old masonry fell on them from unreinforced brick walls.

The earthquake hazards in Santa Cruz had been discussed repeatedly by elected officials and city planners since at least the early 1970s. After the San Fernando earthquake of 1971, the State of California ordered all cities and counties to include seismic safety as part of their planning. By 1975, Santa Cruz planners knew officially that their downtown was particularly vulnerable to shaking in an earthquake because it was built on the soft sediment of the floodplain of the San Lorenzo River, it was near two active faults, and many of the buildings were at great risk in any sizeable earthquake. The city's revised general plan of 1976 recognized the potential for ground liquefaction and severe shaking. What happened downtown wasn't really unexpected.

The crunch came when it was time for action. City planners urged owners to improve their buildings and held workshops on ways to finance the construction costs, but few did any seismic work because they said it was too expensive. As late as 1987 a draft ordinance that would have required property owners to reinforce old buildings was opposed by the Downtown Merchants' Association, and the city council decided in closed session not to pursue the matter.

"We did not see the sense of passing an ordinance without putting the funding sources in place to implement it," Mayor Mardi Wormhoudt said. "We were not ignoring this. We have all been worried about this for years, but the real issue is, how do you put money together to do it? I'm devastated that three people were killed. But if you are asking me if I have second thoughts about our approach, the truth is I don't."

When the city decided not to order earthquake safety improvements to older buildings, the decisions were left to the property owners. Some owners did the necessary repairs, but others said they simply did not have the money. The two old brick buildings responsible for the three deaths in Santa Cruz were among the oldest structures on Pacific Avenue, dating from the turn of the century. Their decorative facades were important to people with a sense of the history of Santa Cruz. Neither had been reinforced, though the owners knew their structures were at risk. Each said afterward that they could not afford the safety work. "No one wants to dig into their pocket to put money into an old building unless they have to," said one of them.

A local civil engineer experienced in the renovation of historic buildings said lack of money is no excuse for a hazardous building. "It was like the owner of the building was driving a car that didn't have brakes," he said of the two buildings that collapsed.

Don Nasser strengthened the building he owns on the mall, remodeling the top two floors of the old Alexander Hotel, a three-story structure similar to the Plaza Books building. He said about $50,000 in seismic work included in the remodeling kept the hotel standing throughout the quake. "Very economical," he said of the work. "I'm sure glad we did it. It saved our entire investment."

Many earthquake experts hope that the devastation on the Pacific Garden Mall will push other cities to address the danger of some their unreinforced buildings. Santa Cruz is not alone: in 1990, more than 300 California cities had about 50,000 old unreinforced masonry buildings, and only 35 of them had set up any kind of plan for seismic reinforcement.

"We must act now or we'll lose the chance," said Gary Griggs, professor of earth sciences at UC Santa Cruz. "Three years from now, society will be worrying about the next crisis, whether it be medflies or AIDS. There are blocks in Hayward, Oakland and San Francisco that will be the next Santa Cruz if we don't do something.

Miscellanea on the Loma Prieta Earthquake.‹The Loma Prieta segment of the San Andreas Fault had been recognized as a particularly earthquake-prone segment in 1983, and in 1988 was assigned the highest probability north of the Tehachapis. However, it was rated as only 30% probability in the following 30 years, so this is hardly a triumph of prediction.

The principal damage was far from the epicenter because of construction on bay fill, in the Marina District of San Francisco, along the section of I­880 that collapsed, at Oakland Airport and Alameda Naval Air Station, and at Moss Landing. The potential problem of the liquefaction of bay fill had been documented time and time again, in particular in San Francisco after the earthquakes of 1865, 1868, and 1906.

The Marina District was built on fill brought in for the site of the Panama-Pacific Exposition that celebrated the opening of the Panama Canal and the recovery from the 1906 earthquake! The fill failed massively in 1989, as might have been expected. The Marina fire of 1989 resulted from a broken gas main, just as the major fires after the 1906 earthquake began in ruptured gas mains in areas built on fill. Just as before, attempts to deal with the Marina fire were hampered by broken water mains.

The major damage in Watsonville, Santa Cruz, and Los Gatos was to buildings constructed before adequate earthquake codes. When most of Watsonville had been repaired or reconstructed, St. Patrick's Catholic church, built of red brick in 1903, still lay in ruins. It was scheduled for demolition at the end of 1992, since it would cost more to repair than to rebuild.

The earthquake and its aftershocks occurred in an area that had been a seismic gap along the San Andreas Fault for the previous 20 years. There were no particular precursors or foreshocks. The earthquake offset along the subsurface fault plane was at most three meters, and there was no major surface faulting. The Santa Cruz mountains are landslide-prone, and landslides were the dominant earthquake-generated damage in the area.

It's clear that this earthquake damaged only the most vulnerable areas. Any larger one will be even more devastating.

Earthquake Prediction

The plain fact is that earthquake prediction has so far been unsuccessful. Predictions have been made in three different ways.

1. On the assumption of seismic gaps: that is, if a sector on a long fault zone that is known to be active has not had seismic activity for a long time, then an earthquake is more likely to occur on that segment. Seismic gaps were proposed as valid indicators of earthquakes about 20 years ago, but they have not turned out to be reliable predictors.

2. Increased activity on a fault has been interpreted as foreshadowing a larger earthquake, as, for example, at Long Valley Caldera.

3. On the assumption that earthquakes occur on a regular cycle, predictions have been made for on-time recurrences.

Bay Area Earthquake Prognosis Here are the odds of major earthquakes (M7 or more) along the major faults of the San Francisco Bay Area in the next 30 years, as follows: €The Hayward Fault north of Fremont to San Pablo Bay: about 30%. €The Hayward Fault south of Fremont: about 25%. €The San Andreas Fault from the Santa Cruz­San Jose highway 17 to Crystal Springs Reservoir just south of San Francisco: about 25%. €The Rodgers Creek Fault, essentially the Hayward Fault's continuation north of the Sacramento River to the Santa Rosa Valley: about 25%. €The San Andreas Fault north of San Francisco; unlikely: maybe 2%. By the quirks of probability, these add up to 67% chance of a major quake somewhere in the Bay Area in the next 30 years.

Scenarios for a M8.3 Earthquake on the San Andreas Fault

These scenarios are official in the sense that they were produced by the California Division of Mines and Geology. The Division was asked to prepare an estimate of a repetition of the earthquake that devastated the Bay Area in 1906 (the "San Francisco Earthquake", and a repetition of the earthquake that devastated much of southern California in 1857 (the "Fort Tejon Earthquake"). Each earthquake was estimated at about M8.3, with shaking lasting for 30­50 seconds.

San Francisco Bay Area.‹The area particularly at risk are those that will receive very strong ground shaking. This includes most of the low-lying ground around San Francisco and San Pablo Bays, and most of the major population centers in the area: San Francisco, Oakland and virtually all the other cities around the Bay, plus the Santa Clara Valley and the Santa Rosa Valley. Ground failure‹landsliding, liquefaction, and open ground cracking‹will occur in many areas of the Bay cities, including Oakland, San Francisco, Richmond, San Jose, and Alameda. Most other areas near the Fault will also experience severe ground shaking.

Many of the freeways will fail because they are built on soft ground around the Bay. Elsewhere, landslides will block freeways, especially on the San Francisco Peninsula. For example, we can anticipate gigantic traffic jams on a badly damaged US 101 from Novato to San Jose, and on Route 17 from Richmond to San Jose. The Bay bridges will survive the earthquake, but will be unusable since their approach roads will fail. Road travel in all the cities will be difficult because of fires, rubble, or restrictions that will allow only emergency traffic to use them.

All rail transport in and out of the Bay Area will be out of action for at least 72 hours. From the south, rail access will be cut where it was in 1906, east of Watsonville at the Pajaro River bridge. Lines from the San Joaquin Valley will be cut at Niles Canyon, east of Fremont, and near Port Chicago. The line from the northwest will fail by ground failure across Suisun Marsh. Railheads at Vallejo and Benicia will be vital transshipment points for supply lines by barge. BART and local lines will also be put out of action.

All the major airports will be put out of action, and the smaller ones at San Jose and Hayward will only have limited capability. The nearest logical center for major airport operations is Travis Air Force Base at Fairfield. The main way in which air support can be helpful after the earthquake is by massive use of helicopters. The most significant access to the Bay Area after an earthquake, therefore, for dealing with casualty evacuation, foodstuffs and other vital supplies, will be by boats‹barges and ferries‹and helicopters.

There are no clear scenarios on water supply. There will probably be serious damage to the aqueducts that bring water to the Bay Area, though the storage in local reservoirs will be enough to maintain supplies. There will probably be extensive damage to the distribution systems, and water is likely to be contaminated. Sewage systems will leak contaminated water into the Bay.

At least 50% of the power-generating plants will be out of action, and there will be power outages even where supply lines are intact. Repairs will be very slow. Any important service or administrative unit (hospitals, fire and police stations, water supply points) will require its own generator. Gas supplies will be practically destroyed, and will pose a major fire risk.

The petroleum refineries in the area are all built on low-lying areas, and are likely to be put out of commission. Pipelines also cross marshy areas, and are likely to break. Proper operation of helicopters will be handicapped without refuelling points within the damaged area.

Los Angeles area.‹The intensity of the M8.3 quake varies according to substrate in the same way as the Bay Area. The main San Andreas Fault runs round the northeast of the Los Angeles Basin, and that means the major damage will be in that area. However, much of the low-lying area near the port, Long Beach, and Huntington Beach are liable to greater damage than one would expect because they lie on softer sediments.

Segments of the freeway system in the northern areas will be closed, and given the usual saturation, that means the system will generally be inoperable. Interstate 5 across the Grapevine will be closed, and so will I­15 through Cajon Pass. US 101 from the north, and I­5 from San Diego will be the only major road arteries open, after minor damage has been cleared. Road travel in all the cities will be difficult because of fires, rubble, or restrictions that will allow only emergency traffic to use them. All railroads into the area will be cut, except the lines from San Diego. Some of the lines within the Los Angeles Basin will be operable.

The airports will have runways that can accommodate emergency supply aircraft, even if they have to land without ground control or refuelling facilities. Overall, the main routes for goods in and out of the basin will be by sea and by air, with handling facilities and control structures largely out of action. The harbor areas will have damage from liquefaction, especially to rail systems: there will be an unknown amount of damage from ruptured oil pipelines that may slow down emergency use of the harbors

Communications will be badly affected: telephone systems will be put out of action or saturated, and microwave relay stations will be destroyed. Perhaps half the radio stations will be put off the air. Emergency services will have to rely mainly on radio rather than telephone systems (until their batteries run out), and repairs will be slow.

Apart from immediate casualties, the greatest threat to the area is the disruption of water supplies. The Los Angeles Aqueduct and the California Aqueduct from the north will both be broken along the San Andreas Fault, and will not be repaired for 3­6 months. The Los Angeles Aqueduct from the Owens Valley crosses the San Andreas Fault 250 feet below ground in the Elizabeth Tunnel, and damage repair will be difficult and hazardous in the face of probable aftershocks: it might require virtual reconstruction of the tunnel. Until then, Los Angeles will have to rely on water from existing reservoirs (at least several weeks' supply). Only the Colorado River Aqueduct will be available for re-supply (the planners assumed that the break on the San Andreas Fault would not extend southeast to the point where the Colorado River Aqueduct crosses the Fault). There will inevitably be thousands of breaks in the water-supply network, and water valves will have to be closed off to prevent massive leakage. One of the major problems is likely to be shortage of water for fire-fighting. In the same way, thousands of breaks in waste pipes will cause sewage leaks that will pollute most waterways, rivers, harbors, and beaches, and will pose a significant health hazard.

Normal water supply and sewage treatment will be badly affected in any case because of power failures. Probably only about 50% of normal power supply will be available: one-third of normal Los Angeles power is imported on transmission lines that will fail, and the local hydroelectric plants on the California Aqueduct near Castaic will be closed down for lack of water even if they are not damaged. At least 50% of the local power-generating plants will be out of action, and there will be power outages even where supply lines are intact. Repairs will be very slow. Any important service or administrative units (hospitals, fire and police stations, water supply points) will need their own generators. Moderate problems only will result from natural gas supplies, since there is a lot of storage within the area. Petroleum supplies will not be catastrophically affected.

Altogether, the Los Angeles scenario is considerably less threatening than that for the Bay Area, though it is still extraordinarily serious for a great many people. The casualty list will vary greatly depending on the time of day the earthquake occurs, but there may be 10,000 dead and 25,000 injured people who need hospital care. Although we can write of scenarios like this in reasonably objective terms, they are likely to be truly catastrophic if the emergency systems are overwhelmed. There is no doubt that this is a clear possibility unless both the hardware and the personnel are prepared for such an emergency.

Alaska 1964

The 1964 Alaska earthquake was one of the largest ever recorded. It was remarkable not only for its severity, but its length: heavy shaking went on for about four and a half minutes. This means that the earthquake contained a tremendous amount of energy, and when the magnitude was re-estimated in these terms, values over M9 are reached. As well as tremendous damage in Anchorage, the largest city in Alaska, the earthquake generated a tsunami that caused massive underwater mudslides in several Alaskan fiords and harbors. Lives were lost at Valdez and Whittier as well as other places in the North Pacific. (In 1958 another Alaskan earthquake triggered a landslide into Lituya Bay, Alaska, and the resultant wave swept 1740 feet high on the side of the Bay.)

The reactions of Alaskans were in keeping with the spirit of the State and the outlook of the inhabitants. It could be called the "pioneer spirit" or the "frontier spirit," but it could also be described as an attitude of "Let me get mine while I can, and to hell with the long-term consequences." It's predictable, then, that it is not easy to recognize in Anchorage today that one of the greatest earthquakes of all time occurred there only 30 years ago. The Anchorage Museum is devoted to Alaskan history and art, but has only one display panel for the earthquake. In general, people do not want to know about it, a common reaction, as we have seen. Here's another anecdote to halp make the point:

At Anchorage International Airport, Concourse C is the oldest part of the terminal building, and houses administration offices and some commuter airlines. In 1993 the airport commissioned a study of the building as part of a new master plan to upgrade the airport. The engineers, an Alaskan form based in Anchorage, reported that the concourse, which had once contained the control tower that collapsed in the 1964 earthquake, was unsafe and should be vacated. The airport immediately hired another engineering firm from California, which recommended only some small-scale renovations. Part of their argument was that the concourse housed comparatively few people at any time, since it served small commuter craft rather than main-line jets. "I'm choosing to believe that occupancy is safe based on the outcome of the second look," said the airport director.

Metropolitan Areas in the USA at Significant Earthquake Risk (RC)

  1. Los Angeles, California (Los Angeles-Anaheim-Riverside, # 2 metropolitan area in the USA, 14 million people): San Andreas system
  2. San Francisco Bay Area, California (San Francisco-Oakland-San Jose, # 4 in USA, 6 million people): San Andreas system. The entire region suffered major damage in 1906, when the population was much smaller and the infrastructure was much less vulnerable.
  3. Seattle-Tacoma, Washington (# 14 in USA, 2.5 million people): subduction zone that has generated massive prehistoric quakes in the area
  4. San Diego, California (# 18 in USA, 2.3 million people): short but prehistorically active faults that cross downtown
  5. Salt Lake City-Ogden, Utah (# 36 in USA, 1 million people): The Wasatch Front behind these cities marks a major active fault line.
  6. Memphis, Tennessee (# 39 in USA, 1 million people). A repetition of the 1811-1812 New Madrid earthquakes would cause massive damage in this region. Memphis is built on a site that is particularly vulnerable.
  7. Charleston, South Carolina (# 72 in USA, 500,000 people). A massive earthquake destroyed much of the city in 188? [
  8. Anchorage, Alaska: (not ranked, 250,000 people): much of the city was destroyed in the great 1964 Alaska earthquake, and much of it is built on glacial sediment that is structurally weak.]

Megalopolitan Areas in the World at Major Earthquake Risk (RC) (Ranking these by size is always difficult because methods of counting their populations vary from country to county. These rankings are by the US Bureau of the Census. Any way you count it, the two largest megalopolitan areas in the world are at VERY HIGH risk of a devastating earthquake.)

  1. Tokyo-Yokohama, Japan. (# 1, 25 million people.) Last major catastrophe: 1923
  2. Mexico City. (# 2, 17 million people.) Last major earthquake disaster: 1985
  3. Los Angeles, USA (# 12, 10 million people.) Very close to the San Andreas Fault.
  4. Cairo, Egypt (# 15, 9 million people). Several major earthquakes over history: last major earthquake in 1754.
  5. Manila, Philippines (# 16, 8.5 million people). Damaging or devastating earthquakes in 1600, 1605, and 1886.
  6. Lima, Peru. 5.5 million people. Close to major subduction zone
  7. Santiago, Chile. 5 million people. Close to major subduction zone
  8. San Francisco, USA. 4 million people. Last major earthquake: 1906
  9. Naples, Italy. 3 million people. Last major earthquakes: 1455 and 1626.

Other major cities at risk, global list: Seattle and San Diego, USA.

Wellington, New Zealand: the nation's capital is built directly on the Alpine Fault, a fault as large as the San Andreas Fault.

Alexandria, Egypt. 3 million people. Destroyed by earthquake and tsunami, around 350 AD.

Bucharest, Romania. 2 million people. Last major earthquake: 1977

Lisbon, Portugal, largely destroyed in 1531 and 1755.

Quito, Ecuador, last major earthquake 1797.

Tokyo, 1923 and Now: a Scenario for Real Paranoids

The Tokyo­Yokohama or Kanto plain has been the most densely populated part of Japan for centuries. The earthquake disaster of 1923 was magnified by bad luck into a national catastrophe. An enormous earthquake (M 8.3) struck just as people were cooking the midday meal on stoves. Fires spread rapidly and could not be stopped. Enormous damage that had resulted from the earthquake was multiplied by a firestorm that raged for two days. At least 140,000 people were killed in collapsing buildings, by the fires, or by drowning as they tried to escape the fire. Damage to roads and railway lines paralyzed the southern part of the country.

Japan has introduced many earthquake protection features into its architecture and its public services. But it is still earthquake country, and that fact cannot be changed. The protection afforded against smaller earthquakes makes the programs worthwhile, but it is not clear at all that the Japanese people are protected against a very large earthquake. One can point to specific weaknesses in their preparedness.

The Tokyo city government moved in 1991 to a new 48-storey building, the city's tallest building. It is designed to serve two functions: as the normal city administration headquarters, and as an earthquake-proof center that will be the center for disaster management in the next large typhoon or earthquake. Television cameras on the roof give real-time, 360° scanning of the city, and microwave transmitters will allow continuous communication with affected areas even if the telephone system is crippled.

Perhaps the most paranoid scenarios of a great Tokyo earthquake are those that involve the worldwide financial effects. The Tokai Bank estimated in 1989 that a repeat of the 1923 Kanto earthquake would cause about $650 billion in damage in the Tokyo area, about 25% of Japan's annual gross national product. A 1994 projection by a Stanford engioneering firm suggested that there would be 40,000 to 60,000 deaths, 80,000 to 100,000 serious injuries, and economic losses totaling between $800 billion to $1.2 trillion. Japanese investors, especially insurance companies, would sell off their foreign holdings to raise the capital for rebuilding. But since Japanese investors provide about 20% of the foreign capital inflow to the United States, the US and probably most of the world would be thrown into an unexpected recession by the financial shockwaves.

Tsunami

Tsunami are waves generated by events that displace a large volume of seawater quickly enough to create a giant shockwave. Most often, they are generated by earthquakes under the ocean floor, though they may also occur as landslides or massive volcanic surges enter water. Tsunami then travel very quickly as very long waves, only to become steep and high as they reach shallow water. As they strike a coastline, they may be very destructive. Some of the greatest earth-related disasters of all time have resulted from tsunami. In 1707 a tsunami killed 30,000 people in Japan, and in 1896 another killed 27,122 people. The Peru tsunami of 1868 killed 25,000.

Tsunami can travel great distances across open ocean without dissipating, and since they move at great speed they can strike distant shores without warning. Surrounded as it is by earthquake zones, the Pacific Ocean is particularly prone to tsunami hazard, and larger Pacific islands have suffered severe damage at times. 100 tsunami have been recorded in Hawaii since 1819, 16 of them causing significant damage. A great tsunami generated by the 1946 Alaska earthquake killed 173 people in Hawaii, and caused $26 million damage to the port of Hilo. The tsunami from the great 1960 Chilean earthquake killed 300 people in Chile, 61 in Hawaii, and 199 in Japan. The 1964 Alaska earthquake set off tsunami that killed 107 people in Alaska, 4 in Depoe Bay, Oregon, and 11 in Crescent City, California. Since 1945, more Americans have been killed in tsunami than have died directly in earthquakes. In Japan, the Okushiri earthquake of 1993 killed 200 people, most of them in the tsunami that followed the earthquake.

Tsunami are well recorded in history. The earthquake that wrecked Lisbon on All Saints Day (1 November) 1755 was followed by a destructive tsunami, and its effects were felt quite severely in the West Indies.

Alaska 1964.‹The massive earthquake of Good Friday, March 27, 1964, wrecked many cities and settlements along the south coast of Alaska. But apart from the damage caused directly by shaking, the effects of tsunami were the most far-reaching and deadly. Tsunami are often molded by the coastal terrain, and the fiords along the southern Alaskan coast often acted to magnify the height and destructive capacity of the tsunami. Thirty-two people were killed by a tsunami that washed over three-quarters of the town of Valdez. 77 fishing boats, about half the fishing fleet of Kodiak Island, were sunk or smashed beyond repair.

Okushiri, Japan, July 12, 1993.‹The island of Okushiri lies just off the southwest corner of the large Japanese island of Hokkaido. It is exposed to waves from three sides, but is a fine base for fishing and tourism.

On July 12, 1993, a M7.8 earthquake occurred under the Sea of Japan close to Okushiri island. The island was badly shaken: some buildings collapsed, killing or trapping several people, and upset stoves set off fires in various communities on the island. However, the greatest damage and loss of life occurred when tsunami swept the shores of the island after the earthquake. There were two major components to the tragedy of Okushiri. First, it lay close to the epicenter of the earthquake. The tsunami arrived within 2 to 5 minutes of the initial shock, before there was time to issue a public warning or to organize any evacuation. Second, the shock occurred at night, 22:17 local time, when people were inside their homes getting ready for bed. All light and power went off as the tsunami struck the main electrical power station at 22:23.

The exposed west coast of Okushiri was swept by huge tsunami, some of the largest ever recorded in Japan. Monai and Inaho, two villages facing the Sea of Japan, were completely devastated by a tsunami that ran up on shore to a height of at least 10 meters. Aonae, a fishing port of 1600 people on the southern tip of the island, was the largest community on the west coast of the island. The southern section of town, on the tip of the island, was completely overswept by two great waves. The first wave, apparently from the northeast, swept over the southern tip of the island and into the houses by the harbor, setting two fishing boats on fire. Seven minutes later the second, higher wave swept through the harbor, wrecked several rows of houses, and carried the burning boats into the town. The harbor was filled with the wood fragments that had once been homes. The second wave also ripped loose a big propane tank, and the resultant fires burned more than 300 homes in parts of the town that were not devastated by the tsunami. Aonae was on fire at 22:40, less than half an hour after the first shock. A massive breakwater and sea wall had reduced the run-up of the waves, but did not save the town.

Damage was extensive all round Okushiri, and also on the nearby coast of Hokkaido. Boats were damaged all round the Sea of Japan, including the coasts of Korea and Siberia. 198 people are known to have died, two-thirds of them in the tsunami, and most of them on Okushiri. In addition, several dozen people are still recorded as "missing."

Nothing could have been done to prevent this tragedy. A nearby earthquake means that people will not receive any useful tsunami warnings. This is in distinct contrast to the situation in Hawaii.

Hawaii's Tsunami Warning System.‹Tourists in Hawaii may at first be puzzled by prominent loudspeaker systems along rather remote beaches, and by brightly painted steel posts high on the hillsides above the beaches. The loudspeakers are part of a tsunami warning system: the loudspeakers will alert everyone in the area, and the posts mark the height above sealevel at which people will probably be safe from an incoming tsunami.

Hawaii lies, of course, close to the center of the arcs of volcanoes and earthquake zones that together make up the "Pacific Rim of Fire," the tectonically active edges of the Pacific Plate of the Earth's crust. Very large earthquakes can occur anywhere along the edges of the plate, and many of them can set off major displacements on the Pacific Ocean floor, so can generate tsunami. Rapidly crossing the intervening oceanic water, tsunami can, and do, strike Hawaii from any direction.

A history of past tsunami disasters caused the installation of the warning system. The tsunami that really caught the attention of the Hawaiians struck in 1946, when much of the town of Hilo was destroyed. Very sensibly, the most exposed parts of town were not rebuilt, but were turned into public parks. Even so, people have been killed in Hawaii in subsequent tsunami, notably the 1959 tsunami that originated off the coast of South America, and the warning system is fine-tuned constantly so that it is perceived as reliable and effective. Similar systems have been established in other countries, too.

Shinkasen, the Bullet Train

Japan Railways designed the Shinkansen, the "Bullet Trains" and introduced them in 1964. They travel at speeds up to 150 mph, and by 1981 the system was carrying an average of 340,000 passengers a day, with a peak of 1 million per day. Each train can carry 2000 people, and 250 trains are operated at once in rush hours. The trains run through difficult terrain and very seasonal weather. Shinkansen travel through mountains, under the ocean, and across densely populated cities. They use the world's longest passenger tunnel. Clearly, the system is vulnerable to the shaking that would be generated by a major earthquake, and an earthquake mitigation system was introduced in 1965.

The Tokaido line (Tokyo­Osaka) has 25 seismic stations, 20 km apart. Each station is wired to local and general headquarters and to the trains. If a station records a horizontal acceleration of 0.04 g (= about Mercalli VI), power is shut off to the track for 10 km each side of the seismometer. The train is automatically braked if the power goes off, and at full speed a train can be stopped in 70 seconds (about 2.5 km). In most cases, the spacing of the seismic stations ensures that a train will not run into a section of damaged track. Seismic shaking in itself is not likely to derail a train, stopped or travelling: a train will derail because the track itself fails.

After the signal is received at regional headquarters, and the train is stopped, the regional staff can decide, on the basis of accumulated information, whether to turn the power back on (for full-speed or reduced-speed travel), or wait for further information, perhaps a track inspection.

Normally P-waves do not trigger the seismometers, because they respond to horizontal acceleration. This means that the S-wave has already arrived when the warning is given. For the bullet train routes that travel north-south up the east coast of Japan (parallel with the Japan Trench), additional seismometers are set right on the coast. This gives the system an additional 15 seconds warning.

The warning system triggered about 5 times a year in the first 20 years. The tracks were bent only twice, which suggested that perhaps the system's sensitivity has been set too high. The system was re-tuned, and currently is running at about 2 stoppages a year. There are also false alarms caused by equipment malfunction or failure, but they are running at one per 2 years. The Japanese are willing to tolerate this level of false alarms, partly because the Shinkansen have such a safety record in any case: there have never been deaths or injuries to passengers in more than 30 years of operation. It is interesting that no other agency is using a similar system in Japan (or anywhere else).