Large-scale tsunami breakwaters are present along the Sanriku coast. They were constructed to protect cities from future tsunamis, because of the region’s long history of devastating tsunamis. The tsunami breakwaters were designed to resist tsunamis that are similar in strength to the 1896 Meiji Sanriku tsunami. Two well-known tsunami breakwaters are located in Kamaishi city and Ofunato city. In Kamaishi, the tsunami breakwaters were constructed at the entrance to the bay; they are 63 m deep and hold the Guinness world record for the deepest breakwaters (Fig. 3, left). Construction of the breakwaters was completed in 2009; they have a 300 m opening and are 670 and 990 m long (Kamaishi port office, 2011). The two tsunami breakwaters in Ofunato city were constructed after the city was struck by a large tsunami with long-period waves caused by resonance with the tsunami generated by the 1960 Chilean earthquake (Fig. 3, right). The two breakwaters are located at the bay entrance where the water is 38 m deep; they have a 200 m wide opening and are 290 and 250 m long (Kamaishi port office, 2011). Construction of the breakwaters was completed in 1967 and successfully protected the city from the Tokachi-oki tsunami in 1968.
However, the 2011 Tohoku tsunami was higher than the designers expected. The tsunami caused major damage to the breakwaters and inundated both cities. Nevertheless, the breakwaters helped to reduce the impact of the tsunami (both tsunami height and arrival time) on the cities, especially Kamaishi, where many houses still remain (Fig. 4). Figures 5 and 6 show the performance of the Kamaishi breakwaters and the mechanisms by which they were damaged (PARI, 2011). The breakwaters were located on a rock foundation. Thirty-meter-wide blocks were arranged on top of the rock foundation along the direction of the axis of the breakwaters. The blocks rose 6 m above sea level and were designed to protect the city from a 5.6 m high tsunami. A tsunami height of 6.7 m was measured at a GPS station in Kamaishi Sea. On the basis of these data, two simulations were performed for cases with and without breakwaters (PARI, 2011). From the results, the height (mean sea level, MSL) of the tsunami was 10.8 m in front of the blocks and 2.6 m behind the blocks; therefore, the blocks helped to reduce the tsunami height by 8.2 m (Figs. 5, 6). With regard to inundation by the tsunami, the breakwaters reduced the tsunami height (at the shoreline) from 13.7 to 8.0 m and reduced the runup height from 20.2 to 10.0 m (PARI, 2011). Because of the strong current in the 30 cm spaces between the blocks, the rock foundation was damaged. Eventually, ~70 % of the blocks were destroyed. This process occurred slowly; as a result, the arrival time of the tsunami inundation was delayed by 8 min (from 28 to 36 min) (PARI, 2011). However, the tsunami breakwaters at Ofunato were more seriously damaged and are currently submerged in the sea. Possible reasons are that the Ofunato breakwaters were constructed using earthquake resistance design of nearly 40–50 years ago and the wave period of the strong tsunami current might have been nearly the same as the natural period of a wave inside Ofunato bay.
Seawalls are found almost everywhere along the coasts of Japan. According to reports from Ministry of Land, Infrastructure, Transport, and Tourism (MLIT, 2011), the length of the seawalls damaged and destroyed in Iwate, Miyagi, and Fukushima prefectures is ~190 km out of a total length of ~300 km. According to the reports, tsunami overflows of <1 m caused a relatively small amount of damage but overflows larger than 3–4 m completely destroyed the seawalls because most of them were designed to protect the land from high tides or typhoons. However, some of them, for example the seawall in Taro town, were meant to serve as tsunami barriers. Taro town experienced tsunamis in 1611, 1896 (a tsunami height of 15 m, 83 % fatality, and 100 % of the houses destroyed) and 1933 (a tsunami height of 10 m, 32 % fatality, and 63 % of the houses destroyed). In 1934, construction of two, 10 m high seawalls (measured from the mean seawater level) was started; the purpose of the seawalls was to protect the town by allowing the tsunami to flow along both sides of the seawalls. They were completed in 1958, two years before the 1960 Chile tsunami, and could fully protect the town from a maximum tsunami height of 3.5 m. In the 1970s, the town constructed another two lines of 10 m high seawalls to accommodate the increasing population (Kamaishi port office, 2011). The total length of the seawalls is ~2.4 km, as shown in Fig. 7, left. The designs of both of the seawalls took only the 1933 tsunami into consideration. However, the 2011 tsunami flowed over the two-line seawalls, damaged most houses, with 5 % fatality, and destroyed the eastern part of the new seawall (Fig. 7, right).
There are three main reasons why the seawalls were damaged.
The two seawalls crossed in an X shape, which caused the tsunami to accumulate and increase in size at the center of the seawalls.
The foundations of the seawalls were weakened by the river on the eastern side of the town. Soil properties near rivers may have disrupted the stability of foundations.
The seawalls were not maintained properly and had not been adequately connected to each other. The tsunami flowed over the seawalls and became a high-speed water jet. The strong current at high speed caused scouring around the foundations.
Examples of damage to typical seawalls can be found in Ishinomaki city (Fig. 8, left) and in Higashi-Matsushima city. The tsunami height near the control forests of both cities was 7–8 m. On the sea side, the surfaces of the seawalls survived, but on the land side, severe scouring occurred at the foundations. Another example of damaged seawalls is shown in Fig. 8, right. In Yamada town, five blocks of seawalls of total length of 50 m were moved by the tsunami. The block structure survived but failed because of poor connection with the foundations and with neighboring blocks. Figure 9 shows typical mechanisms of damage to seawalls including sliding because of the pressure difference, overturning because of collision of the wavefront, and scouring by strong currents (PARI, 2011).
Fudai village developed along the Fudai River. It suffered from the 1896 and 1993 tsunamis that propagated along the river. In 1984, 15.5 m high tsunami gates were constructed to close the river mouth in case of tsunamis. Fudai was the location of a successful countermeasure structure that protected the village from the 2011 tsunami. The 17 m high tsunami flowed over the gate but inundated only a few hundred meters past the gate (Nikkei newspaper, 2011), as shown in Fig. 10, left. Most of Fudai village, including the evacuation shelters (primary and secondary schools), was protected, as shown in Fig. 10, right, and no loss of human life was reported (Token, 2011). If there had not been a tsunami gate, the tsunami would have damaged the center of the village (Iwate prefecture, 2011).
The residents of Minami-Sanriku town have high tsunami awareness because of previous experience with tsunamis. The maximum height of the tsunami in Minami-Sanriku town was >10 m in some areas, whereas the average height of past tsunamis was <5 m. Seawalls and tsunami gates were constructed at +4.6 m MSL after the 1960 Chile tsunami (Minami-Sanriku town, 2011) and residents did not expect such a large tsunami, because the first tsunami warning had prediced 3 m in Miyagi prefecture. Tsunami evacuation drills are conducted every year. However, the tsunami gates and seawalls were overwhelmed and did not stop the 2011 tsunami, which was higher than 15 m (Fig. 11, left). As a result, 95 % of the town, including the disaster prevention building, was destroyed (Fig. 11, right), and approximately half of the population was missing immediately after the tsunami. Approximately 1,000 people died or are missing as a result of the tsunami.
Another important issue raised by the 2011 tsunami is that many firemen were lost in the call of duty as they closed many tsunami gates and the gates of seawalls. Two-hundred and fifty-four casualties were reported in Iwate, Miyagi, and Fukushima prefectures, and more than 70 of these were while closing these gates (Yomiuri newspaper, 2011a). According to a questionnaire given to 471 firemen in 5 cities (Miyako, Kamaishi, Kesennuma, Ishinomaki, and Iwaki) (Kahoku newspaper, 2011), 61 % of firemen met at their office and went out for duty. Among them, 23 % went to the coast to close the gates, and 47 % went to help the evacuation. The percentages of fireman killed by the tsunami were 22 % during gate-closing work and 31 % during evacuation work. Thus, the Japanese government has a plan to install a new system to control these gates remotely.
An example of a great loss of control forests is in Rikuzentakata city. The city is known for having a 2 km stretch of shoreline lined with ~70,000 pine trees (Fig. 12, left). The 2011 tsunami, which was nearly 20 m high, swept away the entire forest; only one 10 m high, 200-year-old tree remains (Fig. 12, right). This surviving tree became a very important symbol of the reconstruction for people in the city. The forest not only could not protect the town but also increased the impact of the tsunami because of floating debris.
In Natori, where Sendai airport is located, a tsunami with a height of 10–12 m, as measured from garbage remaining on trees (Suppasri
et al., 2012b), overturned most of the trees (Fig. 13, left); however, the control forest helped to protect the airport, because the tsunami inundation depth was only 4 m.
Unlike the first two examples, almost all of the pine trees in the control forest in Ishinomaki survived (Fig. 13, right). The forest reduced the destructive power of the tsunami and trapped debris, for example cars, from the water before it entered the city. The trees may have been saved because the height of the tsunami at Ishinomaki was lower (~6 m). The seawall (which was later destroyed) may also have helped protect the trees. Yomiuri newspaper (2011b) reported results based on the estimates from a field survey of tsunami-affected areas conducted by the Forestry and Forest Products Research Institute. Without control forests, it is predicted that a 16 m high tsunami would have inundated 600 m in 18 min with an average velocity of 10 m/s. However, with the control forest, the tsunami arrival time was delayed by 6 min, and its velocity was reduced to 2 m/s.
In general, control forests can withstand tsunamis up to 3–5 m high, on the basis of historical Japanese tsunami data in 43 locations, namely, 1896 Meiji-Sanriku, 1933 Showa-Sanriku, 1946 Nankai, 1960 Chile, and 1983 Japan Sea, as shown in Fig. 14, left and right (Shuto, 1985). The circles indicate trees that have survived whereas triangles and the rectangles indicate trees that collapsed or were cut down, respectively. For example, a tree with a diameter of 10 cm can withstand a tsunami inundation depth up to 3 m but will collapse or be cut down if the inundation depth is greater than 4 and 5 m, respectively. Figure 14, right, shows the effectiveness of the control forest in trapping debris and reducing the wave current. The effectiveness of the control forest was limited at an inundation depth of 3 m for a forest width of <20 m. Historical data show that a forest width >100 m is expected to be effective up to an inundation depth of 5 m. The maximum 2011 tsunami heights in both Rikuzen-Takata (150 m forest width) and Natori (500 m forest width) were >10 m (out of the data range), and caused devastating damage. On the other hand, a 6 m tsunami attacked the control forest in Ishinomaki (150 m forest width); the damage that was caused is shown in Fig. 14, left.
Figure 15 shows a good example of how control forests and breakwaters could have helped to reduce the damage to areas behind them in Ishinomaki city. This figure was created by visual inspection of satellite images, with gray indicating the area of tsunami inundation by the 2011 tsunami, red indicating the areas where houses were washed away, and blue indicating the areas with surviving houses (TEL, 2011). It is very clear that the number of houses washed away in zone B (behind the control forest) is much smaller than that in zone A (without a control forest). Gokon and Koshimura (2012) showed that the probability of a building being washed away in zone C (inside the breakwaters) was ~40 %, whereas in zone D (outside the breakwaters), it was almost as high as 90 %, confirming the 50 % reduction effect, although both areas experienced a maximum tsunami height lower than 7 m.