Printer Friendly

Reduction and rehabilitation of seismic damage to river dikes in Japan.

Seismic failure of river dikes and current trend of its rehabilitation in Japan are summarized. Serious deformation of dikes was mostly brought by soil liquefaction. Stretching type of failure observed where liquefaction took place in shallow depth of ground was recently noticed. Treatment against liquefaction is essential to mitigate earthquake inducing failure of dikes.


Global trends of urbanization to lowland may cause serious damage to society once a river dike fails due to natural hazards such as typhoon, earthquakes, tsunami, and storm surge, as was seen in the Hurricane Katrina case in 2005. As intense rainfall tends to drop in locally limited area due to global warming adding to the recent trend of urbanization in Japan, the importance of dike role is increasing particularly in the urbanized lowland area where population and infrastructure are concentrating.

Not only the rainfall inducing disaster but earthquake inducing disaster can not be forgotten because the seismicity around Japan is recognized to have become an active period (Oike 1995). As the hinterland was highly urbanized as seen in Figure 1, the Yodo river dike damage during the Kobe earthquake in 1995 gave a big threat to the society.


The river dikes have long history in their construction stage, so it is general that the mechanical properties of embankment material and ground conditions are not fully known. This makes it difficult to find out the failure mechanism and governing factors of seismic damage to dikes when it happens. However, the accumulation of damaged experiences is revealing the complex mechanism of seismic failure of river dikes. This paper summarizes the river dike damage recently seen in Japan (Sasaki et. al 2004).


It was known that earthquake inducing failure to river dikes often caused extensive decrease of its function, and in the worst case, the height of the embankment became to 1/4 of original height. Failed sections of dikes were mostly accompanied by longitudinal cracks where step-like discrepancies were often caused.

Such an extensive damage was frequently seen on liquefied ground. It was also known that sand boiling was often observed along toes of dikes resting on old river courses. This experience implies that liquefaction susceptibility of subsoil could be evaluated from geomorphologic information such as the old river courses even when enough boring data are not available.

Table 1 shows recent earthquakes which caused damage to river dikes in Japan. Those cases can be divided into two groups from the rehabilitation point of view. No soil improvement technique had been conducted in the restoration works before 1993. But from the case of Kushiro-oki earthquake, remedial treatment of foundation ground (Oshiki and Sasaki, 2001) was accepted as a part of restoration works by the Ministry of Budgeting. The causative reason and mechanism of seismic failure of dike became to be revealed from the cases after 1993. Two typical damage are as follows.

Although subsoil liquefaction had been known as a main cause of dike damage, but it was not recognized until the case in 1993 that the large deformation of dike is induced by the liquefaction of bottom part of dike in a particular situation.

Figures 2 and 3 illustrate the damaged section of the 8-9 m high dike of the Kushiro River. It was noticed first from this case why the liquefaction of bottom part of dike takes place. The consolidation of highly compressible peat layer caused dike settlement, and this settlement had increased saturated zone in the lower part of the dike which was liquefied during the event (Sasaki et. al 1994).


Figure 4 shows the damaged section of the Shiribeshi-Toshibetsu River dike during the Hokkaido-nansei-oki earthquake in 1994. In this case of damage, it was apparent from the traces of sand boiling nearby the toe of the dike that the large deformation was caused from the liquefaction in the subsoil layer. It should be noted that large amount of depression was observed at its crest as shown in Figure 4. Careful observation during restoration work at this site revealed apparent two slip surfaces which ran from the shoulders inside the failed dike as shown in Figure 5.


It was reported elsewhere (Sasaki et. al 1997) that these slip planes were brought by the stress change in the dike associated with the change of stress at its bottom boundary due to the loss of shear strength of the foundation layer by its liquefaction.

This finding implies that it is necessary to take the deformation of dike into account adding to the large deformation of liquefied subsoil layer when to estimate the seismic settlement of dike during earthquakes. Also it should be noted that stretching type of deformation usually causes longitudinal cracks and depressions of crest.


As was seen in previous chapter, the soil liquefaction is the key to whether or not trigger large deformation to dikes during earthquakes. So it is considered that the prevention of the occurrence of soil liquefaction is the fundamental measure to mitigate dike damage. Most of the seriously damaged sections in the past shown in Table 1 were restored by using remedial treatment method to prevent the liquefaction (Sasaki et. al 2004) as shown in this Table. The remedial treatments to improve foundation ground against liquefaction are compiled elsewhere (Japanese Geotechnical Society 1998).

Comprehensive study revealed that the main cause of the damage at the Torishima section of the Yodo River dike shown in Figure 1 was subsoil liquefaction. The deformation and settlement of the dike was considered to be caused along the process shown in Figure 6 (Sasaki et. al 2004). Therefore this section was restored with conducting Deep Mixing Method for improving the liquefiable 10 m thick sand layer beneath the dike as shown in Figure 7.


The finding gained from the case of the Shiribeshi-Toshibetsu River dike was utilized for the newly constructed Naka-umi dike. Soil profile at this site showed that liquefaction would be easily triggered during an earthquake as shown by FL value in the Figure 8. However it was hesitated to take soil improvement method to avoid the damage for 3 m high dike in the rural area. So the geo-grid was placed at its bottom to prevent stretching type of failure as shown in the Figure 9. Most of the treated length of the dike section performed well without apparent deformation during the Tottoriken-seibu earthquake in 2000 (Sasaki et. al 2004).



As the role of dikes which protect lowland area from flooding becomes more important than the past, so it is essential to raise the resistance of them against causative action of natural hazard. Necessary room for them to be enlarged is limited in urbanized area, therefore existing dikes must be added ductile quality at their locations. In order to achieve this demand effectively, it is necessary to know the weak sections of dike.

To filtering out the weak sections against earthquakes from the long spanned whole dikes, sections which protect so-called zero meter area are firstly to be selected. Then the seismic performance of the dike in the section is to be examined. As the stability of dike is mainly governed by the occurrence of liquefaction, the subsoil conditions of these selected sections are to be examined. If the liquefaction is anticipated, amount of the dike deformation should be estimated. Crest settlement is taken as the measure of dike deformation so that the dike function is to be evaluated. And then remedial measure to reduce the estimated deformation is to be selected.

There are hidden concepts in the background of the selection of zero meter area as a primary step of the procedure, those are based on the past experiences in Japan that the damaged dike can be tentatively rehabilitated by using soils as its construction material in comparatively short period after the earthquake, and that flooding does not occur at the same time with earthquake.

The most important thing from the geotechnical point of view is the estimation of the seismically induced deformation of dikes. For this purpose, conventional stability analysis by circular arc method has been used for a long time. In this empirical method, both safety factor against inertia force without reducing soil strength, and safety factor against reduced soil strength due to the liquefaction induced pore water pressure in the liquefiable layer where the inertia force is not taken into consideration are separately examined. And the obtained safety factor, smaller than the other, is converted to crest settlement by using empirical relation between the crest settlement and the safety factor.

However it is being disclosed that the deformation mode of the liquefied layer is different from what is assumed in the conventional stability analysis. Therefore revision is being made for estimating the seismically induced crest settlement. Adequacy of four numerical analyses was examined. Those numerical analyses are: a computer code named ALID for static analysis using softened soil concept: the Towhata method, a static approach using the viscous liquid concept: a computer code named LIQCA using coupled effective analysis, and a computer code named FLIP using an uncoupled analysis (Japan Institute of Construction Engineering 2002).

Computed deformations were compared with the dike settlements observed during the Hokkaido-nansei-oki earthquake and the Kobe earthquake. Figure 10 shows the result of this comparison. As seen in this figure, the estimated settlements by analytical approach agree fairly well, but the empirical method predicts always larger settlements than the observed ones. This too conservative estimation by the empirical method arises from the used relationship between the safety factor and the settlement. It is also noticed that the static approach by Towhata method can estimate the dike settlement well as the dynamic approaches (LIQCA and FLIP) do.



The damage to river dikes due to past earthquakes in Japan and their rehabilitation works were summarized. It was shown that the soil liquefaction was the main cause of seismic failure of dikes.

Although the seismic stability of existing dike is already diagnosed, but strengthening is delayed due to limitation of budget. In order to reduce the necessary cost by raising the precision of diagnosis, an adequate method to evaluate the seismic deformation of dike is desired to be established as soon as possible so that the effective strengthening is conducted for the necessary section to mitigate the extensive disaster.


Japanese Geotechnical Society Ed. (1998). Remedial Measures against Soil Liquefaction from investigation and design to implementation, A. A. Balkema.

Japan Institute of Construction Engineering (2002). Analytical methods to predict seismically induced dike deformation, JICE Report No. 102001. (in Japanese)

Oike, K. (1995). "Seismic islands set in an active stage", Iwanami library of science 33, Iwanami. (in Japanese)

Oshiki, H. and Sasaki, Y. (2001). "Restoration works of seismically damaged river dikes using remedial treatment of liquefiable layer." Journal of construction management and engineering, Japanese Society of Civil Engineer, No. 686/VI-52, 15-29. (in Japanese)

Sasaki, Y., Oshiki, H. and Nishikawa, J. (1993), "Embankment failure caused by the Kushiro-oki earthquake of January 15, 1993." Performance of Ground and Soil Structure during Earthquakes, 13th International Conference on Soil Mechanics and Foundation Engineering, New Delhi, JSSMFE, 61-68.

Sasaki, Y., Moriwaki, T. and Ohbayashi, J. (1997). "Deformation process of an embankment resting on a liquefiable soil layer." Deformation and progressive failure in geomechanics, Proc. IS-Nagoya'97, 553-558.

Sasaki, Y., Kano, S. and Matsuo, O. (2004). "Research and practices on remedial measures for river dikes against soil liquefaction." Journal of Japan Association for Earthquake Engineering, Vol.4, No.3 (Special Issue), 2004.


Executive y Adviser, Japan Institute of Construction Engineering 3-12-1 Toranomon, Tokyo 105-0001, Japan
Table 1. Recent earthquakes which caused damage to river dikes

Name of earthquakes M Date Damaged dikes

'68 Tokachioki 7.9 1968/5/16 Tokachi River, Mabechi River
Nihonkai-chubu 7.7 1983/5/26 Hachirogata
Kushiro-oki 7.8 1993/1/15 Tokachi River, Kushiro River
Hokkaido-nansei-oki 7.8 1993/7/12 Shiribesfti-Toshibetsu River
Hokkaido-tolto-oki 8.2 1994/10/4 Kushiro River
Hyogoken-nanbu 7.3 1995/1/17 Yodo River
Tottoriken-seibu 7.3 2000/10/6 Naka-umi Lake
Miyagiken 6.2 2003/7/26 Naruse River
Tokachi-oki 8.0 2003/9/26 Tokachi River
Niigataken-chuetsu 6.8 2004/10/23 Shinano River

Name of earthquakes M Restoration

'68 Tokachioki 7.9 -
Nihonkai-chubu 7.7 CW, Drain
Kushiro-oki 7.8 SCP, SCP
Hokkaido-nansei-oki 7.8 SCP, DC
Hokkaido-tolto-oki 8.2 -
Hyogoken-nanbu 7.3 DMM, SSP
Tottoriken-seibu 7.3 SSP
Miyagiken 6.2 SCP
Tokachi-oki 8.0 SCP, JG
Niigataken-chuetsu 6.8 SCP, GG
COPYRIGHT 2005 World Scientific Publishing Co. Pte Ltd.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2005 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Sasaki, Yasushi
Publication:Geotechnical Engineering for Disaster Mitigation and Rehabilitation
Geographic Code:9JAPA
Date:Jan 1, 2005
Previous Article:Applications of two costal pretection structure construction methods.
Next Article:Zoning for liquefaction potential of Taipei Basin, Taiwan.

Related Articles
State Measures 15 and 16 - Seismic rehab bonds - Yes.
Extreme events in geohazards in Asia.
Natural disaster management in the Republic of Korea.
Geotechnical engineering aspects related to Aceh's tsunami and earthquake disaster and the need for its mitigation strategy.
Zoning for liquefaction and actual damage during the 2005 Fukuokaken-seiho-oki earthquake.
The disaster of residential land fill caused by earthquakes.
Improving seismic response of fine grained soils.

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters