PSI - Issue 54

Koji Uenishi et al. / Procedia Structural Integrity 54 (2024) 67–74 Uenishi / Structural Integrity Procedia 00 (2023) 000–000

69

3

direction is weaker than that in the horizontal direction in the far field. Historically, therefore, structures, both in the

underground and on the surface, have been designed to withstand the dominant horizontal vibration under the

assumption that the vertical vibration and its effect on structures are negligibly small.

In addition to the above abnormally long horizontal shaking, another exceptional or distinctive shaking is the

abrupt vertical shock, really sudden “shock” felt before ensuing other horizontal and vertical shaking. The strong

vertical shock arises from a shallow earthquake, even with a relatively smaller magnitude, that occurs just

underneath. The shock has been verbally reported repeatedly for relatively shallow quakes at many different places

in Japan near their epicenters, for instance, in the Nikko area of Tochigi on February 25, 2013, for an earthquake

= 6.3. According to the daily Japanese newspaper, Mainichi

with a very shallow focal depth of 3 km and M j

Shimbun, issued on the following day, the people in the area with the epicentral distance of less than 10 km were hit,

with a thump sound, by a violent vertical jolt that pushed them up from below. However, while the shaking of long

duration by an interlocking earthquake seems to have been recorded relatively adequately (i.e. similar to what we

feel) as shown in Fig. 1(b), the vertical shock arising directly from below has not been historically “captured” by

seismographs nor “visible” in seismograms. Until now, the existence of such shock has been inferred only indirectly

from the unique dynamic structural failures caused by the related seismic waves in the underground and on the

surface. This contribution describes the difficulty and some recent progress in detecting such shock possibly

containing vibrations of higher frequencies, by introducing actual examples of seismic events attracting great

attention and those perhaps drawing much less caution.

2. Shaken, damaged but no signal detected

As mentioned above, horizontal shaking is normally dominant during earthquakes of short/long duration.

According to the seismological records shown in Fig. 2, this idea holds for the extremely devastating earthquake, the

1995 Hyogo-ken Nanbu (Hyogo Prefecture South) earthquake that directly hit Kobe City and its neighboring

= 7.3. However, dissimilar to “usual”

densely populated urban area, with a shallow focal depth of 16 km and M j

earthquakes where the epicentral distances are of the order of hundreds of kilometers or more, the Kobe area was

struck directly by the strong seismic waves propagating without large attenuation. As a result, some 640,000

buildings and houses on the surface were fully or partially damaged. There were also verbal evidences of

experiencing vertical shock at the earliest stage of seismic shaking, for instance, “There was a tremendous sound, as

if a sudden storm were to create a ‘gurgle’ and knock down the trees in the forest. This was followed by some

‘splitting’ sound as if a large tree were being torn in half, and then a strong impact that seemed to push me up from

the floor to the ceiling” (Jojima, 2009). This evidence implies that, like the waves caused due to blasting by

detonating explosives, the frequencies of the waves generated by this earthquake, possibly longitudinal (P) ones that

are equivalent to sound waves in the air, were at an audible level. According to the International Electrotechnical

Commission (1994), audible frequency for human beings is over 16 Hz, but Fig. 2 shows that even in the epicentral

region, only 16.5 km away from the epicenter of the quake, there exists no evidence of strong vertical shock or

vertical vibrations with higher frequencies. Apart from this, something that should not be overlooked occurred. The

photograph taken near the Sannomiya railway station (Fig. 3(a)) shows surprising structural damage to a reinforced

concrete (RC) building. At a glance, the RC building does not look affected by the quake at all. However, carefully

observed, the photograph shows that only the middle floor was in compression and totally flattened. Astonishingly,

the signboards were perfectly undamaged and the floors above vertically shifted almost without horizontal

movement and deformation (see also Uenishi (2021) for another similar example). Apparently, it seems impossible

for the horizontal shaking to cause this localized vertical compression without horizontal movement.

Not only on the surface but also in the underground something peculiar happened in 1995, although underground

structures were and still are believed to be resilient against earthquakes. In downtown Kobe, the Daikai railway

station constructed by cut-and-cover methods disastrously failed (Fig. 3(b)). At this underground station some 5 m

below surface, the RC central columns supporting the roof vertically collapsed and caused the failure of the roof and

the sinking of the prefectural road above. The important feature of the failure here is that the damage was only

limited to the central columns with specific cross-sections. Only little damage was caused to the columns with small

overburden, and the columns between the stations were hardly damaged. Besides this amazing structural failure, at

the central section of the some 1,750m-long Bantaki road tunnel that had been built by the New Austrian tunneling