A team of U.S. and Japanese researchers conducted a five-day field inspection of the of tsunami borne debris and its effects on structures following March’s magnitude 9 earthquake and ensuing tsunami off the northeast coast of Honshu, Japan.

Field inspections of Sendai, Natori, Ishinomaki, Onagawa, Minamisanriku, Kesennuma, and Rikuzentakata were conducted. The research effort was funded by the U.S. National Science Foundation.

Figure 1: Hydrostatic/hydrodynamic wall failure on seaward face of building.

The observed damage to structures associated with the Tohoku tsunami event can be tied to the hydrodynamic, hydrostatic, buoyancy, and impact demands generated. As the tsunami inundation occurred, the high velocity of the flow resulted in hydrodynamic forces on structures. These demands resulted in flexural failures of reinforced concrete wall structures on the seaward faces of buildings (Figure 1). As the inundation depth increased, structures that were able to remain sealed from the flow were subject to significant hydrostatic demands, resulting in a flexural or bucking failure of the walls. This was observed for both walled building systems with minimal openings and container structures such those present at fuel storage facilities (Figure 2). For structures that were able to remain sealed or structures that had low density, significant buoyant forces were generated, resulting in failure of the connections to the ground. This was observed for fuel storage containers, wood frame structures, and for an autoclaved lightweight concrete structure.

Figure 2: Buckling failure of container structure.

Damage to walled structures was not limited to inundation only. Flexural failures of reinforced concrete walls were also observed on the non-seaward facing portions of buildings. This damage is associated with rundown due to receding water. In many regions examined, the inundation depths were in excess of 5 meters. Due to the long duration and high depth of the inundation, these buildings had time to fill with water through the small openings in the walls. During rundown the depth of water outside the structure was able to decrease at a much faster rate than the water in the building, resulting in a differential hydrostatic demand. This resulted in flexural failure of the walls in an outward direction, causing spalling of the inside face of the building columns indicative of outward pressures (Figure 3).

Figure 3: Outward hydrostatic flexural failure of walls due to rundown.

Observations identified debris consisting of a variety of objects, including wood poles, trees, entire houses, shipping containers, vehicles, small fishing boats, large ships, and small planes. The impact of this debris on structures was found to be sensitive to the relative sizes of the debris and structures, as well as the flow characteristics around or through the structure.

For tsunami borne debris to inflict damage to a structure it must have adequate velocity and mass to generate significant force demand on impact. For the object to have a high velocity, the velocity of water in which it is traveling must be high in the vicinity of the structure. For the velocity to be high at impact the flow of water must be unimpeded. This becomes a relative size scenario. For example, a raised building or bridge on columns would have minimal effect on the water velocity at the structure. Moderately sized debris such as a container or car moving with the flow would likely impact a column or elevated beam as it attempts to pass through the structure. This was observed for a parking structure in Kessennuma that was constructed with an open first floor level (Figure 4).

Figure 4: Impact damage due to high velocity flow at first floor

The same debris traveling toward a walled structure would likely flow with the water around the structure and would at worst impart a glancing impact demand to the building. If the debris has a very large mass and geometry, such as a steel hulled ship, it may not change direction as easily, which would result in significant impact to a similarly sized walled structure (Figure 5).

Figure 5: Ship to building impact.

Debris was also found to settle in many cases during the rundown. Boats, cars, and trees were observed on top or within structures on the inland face of the building. As water recedes the floating debris moving back out to the sea becomes trapped against or on top of structures. As the water level continues to decrease, settlement of the debris results in additional vertical demands on the structure. While this demand can be significant it would likely be applied in a quasi-static rate.

The observation of buildings and building components becoming debris was also significant. Due to the buoyant nature of wood construction and the relatively weak tie-downs used, many wood framed houses detached from their foundation and became floating debris. This type of debris has a high mass and can consequently generate significant impact forces, even at low velocities. Another characteristic of the "debris problem" is that damage can occur above the maximum water line because of the buoyant nature of the debris. For example, damage was observed to a school building on the second floor 1 to 2 meters above the inundation depth due to the buoyant debris from several timber frame houses. Due to the ubiquitous use of heavy ceramic tile roofs the impact demands imparted are significant. As a consequence of this observation, resilient structures would need to consider not only the hydrostatic and hydrodynamic demands below the inundation depth but also the possibility of a simultaneous impact above the expected inundation level.

Another major aspect of the debris is "debris damming effect." Although the study focused on the effects of debris impact, we also noted widespread effects of debris damming where debris such as light metal siding became enmeshed in steel moment frame structures, significantly increasing the drag forces on the structures, leading to failure. In other cases, such as warehouses located directly on the waterfront, debris damming was less significant and the failure of the siding allowed the tsunami to flow easily through the building, preventing collapse of the structure.

Further study and processing of the field observations and measurements are underway to fully characterize the demands and associated damage observed.

The authors would like to thank Professor Mizutani of Nagoya University and Daiki Tsujio of Pacific Consultants Co. for their assistance during the reconnaissance.

Clay Naito, corresponding author, is an associate chair, Department of Civil and Environmental Engineering, Lehigh University. Contact him at cjn3@lehigh.edu. Daniel Cox is a professor at the School of Civil and Construction Engineering, Oregon State University, Corvallis, Ore. Qi-Song "Kent" Yu is a principal at Degenkolb Engineers in Portland, Ore.