A tsunami is a series of waves generated by underwater earthquakes, volcanic eruptions, or landslides. As these waves approach the shore they lose their speed but greatly increase their height, and all coastlines are at risk from them. When a tsunami is the result of locally occurring phenomena, the resulting waves strike almost immediately and are particularly devastating. In July of 1998, a moderate earthquake near Papua, New Guinea caused a tsunami with 15m waves that hit the coast a few minutes later, killing at least 2000 people. But tsunami damage also occurs as a result of distant events. During the 1964 Great Alaska Earthquake a tsunami caused damage not only in the Gulf of Alaska, but also along the west coast of North America and Hawaii.

Tsunami waves travel across the deep ocean at 800km/h; a tsunami generated in Alaska would take almost a day to arrive in Antarctica. Although many tsunamis occur in the Pacific, the Atlantic Ocean is also at risk from tsunamis not only from earthquakes, but due to submarine landslides, volcanic activity, and meteor impact.

Because tsunamis are caused by the displacement of water - from earthquakes, undersea volcanos, or submarine landslides - all regions are at some risk. The Canary Islands have submarine canyons with unstable soil capable of creating large tsunamis that could impact the western coast of Europe. The 1755 Lisbon Earthquake had a submarine source in the Atlantic that caused inundation of the coast of Portugal, Spain, and Morocco. Undersea volcanic activity in Iceland could impact low-lying areas in Northern Europe.

But some regions are much more at risk than others, for instance Japan and Hawaii are routinely innundated by tsunamis due to their position in the Pacific Ocean and also due to them having very steep volcanos whose flanks periodically slide into the sea. Regions with unstable submarine canyons inside bays are most at risk.

Over the last 60 years, at least ten earthquake-induced tsunamis occurred in the Pacific Ocean. Many of these events caused bridge damage including collapse. But no bridge owners currently address tsunamis in their design codes.

During a tsunami, large waves roll onshore in what is known as a 'run-up', forcing the water in rivers to flow upstream, and filling bays. The height of the incoming waves can be more than 15m; when they reach the shore, the waves slow to a velocity of about 24km/h. The waves can progress inland for almost a kilometre, and for several kilometres up streams, rivers and bays. The water then flows back, often dragging cars, buildings, and other objects with it.

As it recedes, this water scours the soil away from foundations. Bridge foundations are exposed to about three times the substructure width or even greater if the water is forced to flow under the bridge soffit. This scour removes the support from around the foundations and makes bridges more vulnerable to the static and dynamic forces associated with the tsunami.

The 1960 Chile earthquake had the largest magnitude ever recorded, and the resulting tsunami damaged bridges thousands of miles away. One bridge on Honshu Island in Japan experienced heavy scouring of the channel bottom, causing a pier to settle by some 900mm.

Large dynamic forces acting on bridges during tsunamis can reach ten to 20 times the hydrostatic head of the water. As the wave strikes an object, the slope of the wave front causes a sharp increase in the acceleration of the water resulting in large surge forces on pier walls, columns, and other bridge substructures.

Similarly, the larger the obstruction in the tsunami's path, the greater the drag force on the object. A circular column may have a coefficient of drag less than 0.5 while the coefficient of drag on a rectangular, submerged pier can be up to 2.

During the 1946 Aleutian Island Earthquake, a large tsunami struck Hilo, Hawaii. The water rose 5m at the mouth of the Wailuku River and the steel truss span of a railroad bridge was pushed upstream. At the mouth of the Kolekole River, just north of Hilo, the tsunami destroyed a steel tower and the steel girders of a railway bridge, although the rails remained in place. The railway was never rebuilt.

But the most significant of the dynamic forces acting on bridges is due to the impact of objects striking bridge piers and superstructure. This problem is not restricted to large objects such as boats and cars; tiny objects like soil grains can sandblast bridges.

During the 1964 Great Alaska earthquake, bridges were damaged along the west coast of North America. Damage to the Copalis River Bridge in the State of Washington was the result of the tsunami picking up trees and smashing them into the timber piers.

Two static forces associated with tsunamis can damage bridges. As the water rises, buoyant forces can pull hollow piers from their foundations and lift hollow box superstructures off their bearings. Hydrostatic forces can push over bridge wing walls and abutments. The soil becomes saturated by the tsunami and when the waves recede the increased hydrostatic head knocks the retaining structure down.

During the 1994 Kuril Island earthquake, a tsunami picked up the superstructure of a bridge on Kunashir Island. Frequently, tsunamis reverse the flow of rivers as they travel upstream from the sea. This causes flooding in low-lying areas near the river banks. This bridge approach, located 11.5km north of Yuzhno-Kurilsk on Kunashir Island, was washed out by the tsunami.

Further information about tsunami hazards can be found in the National Tsunami Hazard Mitigation Program report Designing for tsunamis, in the US Corps of Engineers' manual, Tsunami engineering, and from regional reports that provide local and distant tsunami amplitudes based on the shape of the coast, the stability of submarine geology, the location of undersea faults and so on.

Designing new bridges for the large and varied forces associated with tsunamis can be a daunting task. For this reason, the preferred option is to locate bridges and the transportation system they carry either inland or high enough to avoid tsunamis. No bridge can be designed to survive an impact with 100,000t tankers and cargo vessels, so when it is essential that the bridge remains open, the only solution is to relocate it away from the hazard.

When local conditions are capable of producing an earthquake-generated tsunami, strong shaking is likely to have dissipated before the onset of the waves. In such circumstances, it is enough to ensure that the bridge has adequate reserve capacity after ground shaking to resist the tsunami.

The Pan-American Highway was built high enough to avoid inundation from the tsunami in Camana, Peru, during the 2001 Southern Peru earthquake, although structures on the shore below were damaged. However it was still vulnerable to landslides and embankment failure due to ground shaking.

The first step in tsunami design is to determine the performance criteria for the bridge. This establishes the maximum wave height and the amount of damage bridge owners are willing to accept. Some bridge owners may be unwilling to design for the maximum tsunami but may be more willing to design for a tsunami that is likely to occur during the life of the structure.

Sometimes this information can be difficult to obtain, but California is fortunate that the tsunami hazard along its coast has been the subject of a number of studies. In the 1970s and early 1980s James Houston wrote a series of tsunami prediction reports for the west coast of the US. More recently, Costas Synolakis at the University of Southern California has begun creating a series of tsunami hazard maps for the state. Also, California's Division of Mines and Geology has developed tsunami hazards for Humbodlt and Del Norte Counties in California.

Two kinds of bridges can be designed to resist tsunami forces. Submersible bridges allow the tsunami wave, and hence objects carried by the wave, to flow over the deck while elevated bridges are designed to raise the superstructure high enough above the waves to prevent large objects from impacting the deck.

These structures must be designed with foundations deep enough to support the dynamic forces after scour has removed the upper soil. Abutments must be designed to resist hydrostatic forces and the bridge must have elaborate fenders or armour to prevent vulnerable columns or piers being struck by floating objects. The superstructure must be designed to resist buoyant forces and should have streamlined edges to reduce impact and drag forces, although this can be difficult since most bridges must have a crash barrier rail at the edge of the superstructure. In addition, the superstructure should be designed to be able to support the largest possible object that might land on its deck. Alternatively, the bridge may be designed to be tall enough to avoid the tsunami and any objects carried by the waves from impacting the superstructure. Slope paving should be provided at the soil embankments behind the abutments to prevent washout of the soil.

The most important step is for bridge owners to create a policy that requires new bridges along the coast to be provided with tsunami protection.

Reducing the vulnerability of existing bridges is an even more challenging task. The first step is to identify all bridges along the coast or along coastal bays and rivers that will experience partial inundation from a design tsunami based on the performance criteria. California has several hundred bridges that are vulnerable from scour or from complete submersion during the 12m run-up of the design tsunami.

It is difficult to protect bridges by sheltering them from the incoming waves with walls or trees because the obstacles that bridges span - roads, railways, and rivers for example - cannot be obstructed.

One retrofit solution commonly used for bridges on liquefiable soil may provide protection from tsunamis. Existing bent caps are enlarged to support the superstructure on very large diameter cast-in-drilled-hole piles. The piles are designed to resist drag and surge forces from the tsunami along with the resulting scour. Fenders should be provided to protect the bridge from impact, similar to those suggested for new bridges.

Mark Yashinsky works at the Caltrans Office of Earthquake Engineering

Printer friendly version

Email this article to a friend