Japan Sea Central Region Tsunami of May 26, 1983: A Reconnaissance Report (1984)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

3 TSUNAMI PROPAGATION AND COASTAL TRANSFORMATION The major axis of the ellipse of the Japan Sea central region earthquake borders the deeper portion of the Japan Sea known as the Japan Basin. Water depths along the major axis and in the western portion of the ellipse are in the range of 2,500 to 3,000 m, as shown in Figure 2. East of the major axis and toward the coast of Honshu, Japan, the water depths decrease gradually to about l,000 m before encountering the continental slope. At the continental slope the water depths decrease rapidly to approximately 200 m before encountering the continental shelf. From the shoreline the continental shelf varies in width from a few kilometers at Fukaura and Yatsumori to 30 km at Noshiro to a kilometer or so at the Oga Peninsula. Figure 3 shows the detailed bathymetry of the area. These large variations of bathymetry along the eastern Japanese coast profoundly affected the nearshore behavior of the tsunami. The tsunami first arrived at the eastern Japan coast about 7 minutes after the earthquake occurred. The tsunami consisted of three to four main waves, of decreasing amplitude, each with a period of approximately l0 minutes. Figure 4 shows a photographic sequence of the first wave striking the shoreline in the Oga Peninsula area. These photographs also provide accurate time data because the camera that was used displayed the date and time at the bottom of each picture. The first photograph shows the wave as it was first seen offshore at l2:l2 p.m. The remaining two photographs were taken at l2:l3 and l2:l7 p.m., respectively. During this 5-minute interval, the water rose and receded

FIGURE 3 The detailed offshore bathymetry off the Oga Penin- sula (l), Noshiro (2), Yatsumori (3), and Fukaura (4).

FIGURE 4A The first wave approaching the Oga Peninsula area at l2:l2 p.m. Source: T. Sawaki. FIGURE 4B The same area at l2:l3 p.m. Source: T. Sawaki. FIGURE 4C The same area at l2:l7 p.m. Source: T. Sawaki.

to approximately its initial level, corresponding to a complete wave period of about l0 minutes. Offshore of Noshiro the first wave resembled an undular bore, with a short-period undulation of approximately l0 seconds superimposed on the leading portion of the main wave. At other coastal sites the character of the main waves varied, ranging from undular bores, spilling breakers, and plunging breakers to surges at a few locations. Schematic drawings of the observed waves at various coastal sites are shown in Figure 5. Figure 6 presents evidence for the existence of several waves. This sequence of photographs was taken from an elevation of about 300 m on the Oga Peninsula looking north toward Noshiro. In the first photograph FIGURE 5 Schematic drawings of observed wave configurations. The top record is from Yatsumori. The bottom records are from near Iwadate (north of Noshiro). Source: N. Shuto., personal communication,

l0 id 8 V u o CO o> c •H 1 <d C •a « 0. I 2 m M-l o 3 <D •H B H

ll (upper left), three waves are visible: the first very near the shoreline, the second breaking offshore, and the crest of the third farther offshore. In the second and third photographs (upper right and bottom left, respectively), the first wave has reflected from the shoreline and dispersed into a train of undulations. At least l0 wave crests in the reflected wave can be detected in the third photograph. In the last photograph (bottom right), the reflected waves have collided with the incoming second wave to form a sea of "white water." The formation of waves with a period of 10 seconds at the crest of a main wave with a period of approximately l0 minutes was not expected. The Port and Harbour Research Institute, Ministry of Transportation, has reproduced this phenomenon with experiments in a laboratory wave flume. The institute simulated the bathymetry offshore of Noshiro and generated appropriate incoming waves, as shown in Figure 7. The simulation demonstrated the formation of short-period waves on the leading edge of the main wave as the wave propagated over a steep continental slope and onto a gentle continental shelf. Analyzing the behavior of this transformation is crucial to understand tsunami transformation and to design practical coastal protection structures. Another interesting feature observed was that once the bore formed at one location on the propagating wave, it tended to migrate across the entire wave crest. The wave crest was subsequently propagated with remarkable uniformity toward the shoreline. This phenomenon is shown in a photographic sequence of the incoming wave in Figure 8. Another phenomenon that occurred during the tsunami was the creation of an edge bore that propagated along the shoreline and caused significant damage to shore structures. This edge bore is illustrated in Figure 9B. The open-ocean propagation of tsunamis to Japan and other coastal sites bordering the Japan Sea can be accurately predicted by linear long-wave theory. Theoretical predictions for the May 26, l983, tsunami are shown in the refraction diagram of Figure l0, which was prepared by

l2 80 go STATION C 100 t (SJ I 10 20 CS) 30 40 STATION G 60 Stations at which wave records are shown 60 I CS5 FIGURE 7 The experimental arrangement and a sample of the mea- surements at stations G (beginning of continental slope) and C (near shoreline). Source: Port and Harbour Research Institute, Ministry of Transportation. T. Hatori. It should be noted that the Japan Sea remained disturbed by the tsunami for a period of about five hours, indicating multiple reflections of the wave. Wave trapping also occurred at some coastal sites, in particular, wave activity around Oki Island (see Figure l0) persisted for nearly l2 hours after the wave first arrived there, which was approximately 90 minutes after the earthquake occurred. The extent of wave run-up along the Japan coast varied considerably with location. Figure ll shows a bar graph of the run-up along the

l3 •!J mi J55 * 0 « a <? o 0) -u M O O (0 <D 5 •o H id 3 n GO o c 14-1 <0 Ijl c en •H c 4J -H (d x: Cn en <d -H I 1 ll en a) 0) OH M (0 o &i 0) S M 0) o 2 ja •H <u m •^ -^ •P ra M-l -H Ck •• (fl Q) ^ U o tn si Pi oo H 3 tj en OH C o c H 0)

l4 FIGURE 9A Edge bore sweeping along the riverbank at Oga City. Source: Asahi Newspaper Publishing Company, l983. FIGURE 9B Edge bore along the coastline at Hachimori. Source: K. Yamanouchi.

l5 128-E ixr 112- I4O" 142- FIGURE l0 Refraction diagram for the May 26, l983, tsunami. Source: Japanese Tsunami Committee, l983. Japan coast and Sado Island. Clearly, most of the impact occurred north of the Oga Peninsula, where a maximum run-up of l4.9 m was recorded. More detailed results by Nobuo Shuto and his colleagues at Tohoku University demonstrated a significant reduction in run-up for areas protected by breakwaters. This effect is attributed to the presence of

16 Sakata riiigata FIGURE ll Run-up measurements along the Japan coast. Source: Shuto, l983a.

l7 the short-period l0-second waves that appear to have evolved on the main wave during propagation over the shelf. The breakwaters were designed for protection against storm waves and would have been totally ineffective for the tsunami's long-period waves had these short waves not evolved. Significant wave run-up was also observed along the Korean coast. Figure l2 gives a visual recording of run-up and run-down elevations taken in the Port of Imweon by an observer from the Korea Maritime and Port Administration. A maximum run-up of 3.2 m was followed by a run-down of -5.6 m. Most of the impact of the tsunami in South Korea I 0 I -2 -5 -6 1 1 I I I I I I I I I I I 2:15 2:30 2:45 3:00 3:15 TIME FIGURE l2 Visual observations of tsunami elevations at Imweon, South Korea. The run-ups were measured on the wall of a build- ing; the run-downs were measured at the wall of a dock. Source: Korea Maritime and Port Administration.

l8 occurred between Imweon and Sogcho (see Figure l0). The reason for the concentration of the impact in this area is not clear. It may be the result of the focusing of wave energy by refraction as the waves passed over the Yamato Rise. There were'many tide gauges in operation along the eastern coast of Japan that were available to record phenomena associated with the tsunami. However, due to the slow responses of the tide gauges, the slow chart speed of the recorders, and the relatively short wave period of l0 minutes, data from the tide gauges are not very reliable. Sppli: 1<W*^^^ mm j£±Ead: l0 cm FIGURE l3 Tide gauge records. Top: Oga Aquarium. Bottom: Iwanai Harbor.

l9 Nevertheless, for purposes of information two tide gauge records have been included (see Figure l3). The 10-minute wave period can be seen in the top chart of Figure l3. The bottom chart, however, indicates that there were approximately 20 waves within a six-hour interval, which implies periods of about l8 minutes. This longer wave period could be due to the effects of the harbor's response.