To date, according to reports, the failure of well-engineered dams during an earthquake event has not resulted in any fatalities. Although this is encouraging news, such a favourable performance of dams does not guarantee inherent safety against earthquakes.

For example, during the Bhuj earthquake on 26 January 2001 in Gujarat Province in India, about 200 earth dams were damaged and needed repair and/or strengthening. As the water levels were extremely low during the time of the earthquake, there was no catastrophic release of water from the reservoirs of the severely damaged dams.

Despite the fact that these earth dams, which are less than 30m high, are often built by local communities and differ from the well-engineered dams for hydro power projects, we have to recognise that plenty of similar dams exist all over the world. The water stored behind these earth dams is used mainly for irrigation and water supply.

Furthermore, there are very few large dams which have been exposed to ground motions that may be expected during the maximum credible earthquake (MCE); an event which a dam must be able to resist successfully according to the current guidelines prepared by icold‘s Committee on Seismic Aspects of Dam Design.

The Bhuj earthquake also demonstrated that in countries with inadequate earthquake preparedness, strong earthquakes can cause a large number of casualties and huge economic losses. The problem is the largely unknown and often insufficient earthquake safety of existing buildings and infrastructure projects, as earthquake actions may not have been taken into account adequately in their original design.

Although earthquake regulations exist in most countries, if they are followed properly they only apply to new structures. The earthquake safety of many old structures is essentially ignored. Unfortunately, the same applies to existing irrigation dams in most parts of the world. Although it may be known that these relatively small dams do not comply with today’s rigorous design criteria, as long as no catastrophic incident occurs, the owners and dam safety authorities are still reluctant to look into the safety of these structures.

If we ignore the call for the increased safety of existing infrastructure projects with large damage potential – dams belong to this category – the many groups who are already opposing new dams will use earthquake safety as one of their arguments. There is already evidence that this happening.

The majority of older dams were built using methods of seismic analysis and seismic design criteria which, today, are considered as obsolete or outdated. Therefore, in many cases, it is not known if an old dam complies with current seismic safety guidelines published by ICOLD.

At the annual ICOLD meeting in Antalya, Turkey in September 1999, the Committee on Seismic Aspects of Dam Design was given the task to address the issue of the earthquake safety of existing dams. This is also one of the subjects which will be discussed during the forthcoming ICOLD congress in Montreal, Canada in 2003.

According to current ICOLD guidelines, large dams have to be able to withstand the effects of the so-called MCE. This is the strongest ground motion that could occur at a dam site. In practice, the MCE is considered to have a return period of several thousand years (typically 10,000 years in countries of moderate to low seismicity).

Because of the very long return period of destructive earthquakes in many parts of the world, and because relatively few dams have been severely damaged by strong earthquakes, it is rather difficult to convince dam owners and decision-makers of the benefits of a seismic reassessment and upgrading of deficient dams. Risk analyses carried out for several dams in industrialised countries have shown that the failure of a large dam and the resulting flood wave may cause a large number of casualties and huge economic and environmental damage exceeding billions of US dollars. Earthquakes have the potential to cause the failure of dams that have inherent weaknesses but, unfortunately, the latter are often not known. Statistics on dam incidents show that quite a number of deficient dams fail during the first few years after construction.

Over the past few decades, significant progress has been achieved in the assessment of seismic hazards at a dam site and the dynamic analysis of dams. The trend goes towards higher intensities of the earthquake ground motion at dam sites, which is usually characterised by the peak ground acceleration (PGA). To illustrate this problem: most dams were designed against earthquakes using a so-called pseudo-static approach and a PGA of 0.1 g (g = 9.81m/sec2, acceleration due to gravity). An MCE with a magnitude of larger than six can generate (locally) a PGA of more than 0.5 g, ie a value, which is five times larger than the design value. Because of this large discrepancy between the design acceleration and the PGA values to be expected during the MCE, it is often not possible to make a reliable statement about the earthquake safety of an existing dam.

None of the dams damaged by the Bhuj earthquake in India were equipped with strong motion instruments. Therefore the exact level of ground shaking which caused the observed damage is unknown. This would be essential information to judge the vulnerability of similar irrigation dams to strong ground shaking all over the world.

During the devastating earthquakes which hit Turkey and Taiwan in 1999, very few earthquake records of dams that experienced strong ground shaking are available. It was reported that at the 181m high Techi arch dam, where a PGA of 0.87g was recorded below the crest spillway, power failure rendered some of the strong motion instruments at the dam site as being useless. Also the critically damaged Shih-Kang dam was not instrumented, despite the fact that Taiwan has one of the densest and most sophisticated strong motion networks. In a nearby town, a PGA of 0.57g was recorded.

In 1994, during the Northridge earthquake near Los Angeles, US, some of the important records obtained from the well-instrumented Pacoima arch dam, where PGA values exceeding 2g were recorded at the dam crest, could not be used as some of the older strong motion instruments could not accommodate such high accelerations. It was not possible to perform a reliable analysis, which would be needed to determine the damping of the dam during very strong ground shaking. This is a key problem in many seismic analyses of concrete dams that has not yet been resolved.

Furthermore, the 106m high Sefid Rud buttress dam in northwest Iran, which was severely damaged during the 1990 Manjil earthquake, is one of the most important reference cases for dam engineers. It did not have any properly functioning seismic instruments. At the time of the earthquake, the epicenter of this 7.5 event was located within less than 1km of the dam site. The seismograph was being repaired in Tehran.

The main conclusion that can be made is that the earthquake safety of most existing dams is unknown and some may even be unsafe. If a dam should turn out to be unsafe, then the easiest way to comply with safety standards would be to lower the reservoir level or to decommission the dam. Because there are very few viable alternatives to dams in many developing countries, decommissioning or lowering the reservoir level would be the last resort.

Based on our experience with the seismic safety evaluation of dams in countries of high and moderate seismicity, such as Iran and Switzerland respectively, we can state that well-designed dams will satisfy today’s seismic safety criteria.

We feel that it would be appropriate to address this important subject and take adequate action as structural safety is one of the most important aspects of any dam project, way ahead of economic, environmental, ecological and socio-political concerns. This fact may have been overlooked in the recent debate on benefits and concerns of dams. It is also in the interest of the dam community to have a clean record, as the failure of a single large dam may increase opposition against any new dam projects worldwide.

Investing in the seismic safety of dams is generally considered to be a low priority as there is no visible immediate return on the investment. This has been a problem with all projects related to natural disasters with a very low probability of occurrence. Such investment seldomly reaps benefits for those who have taken the decisions. However, a long term view in such programmes is a prerequisite for the sustainable safety improvements of existing dams.

As there is a steadily increasing demand for water, flood protection and clean energy, the safe operation and structural integrity of dams are vital. Unfortunately, the seismic safety of dams and other infrastructure projects is an area which has been neglected. This is despite the fact that the decade 1990 to 2000 was declared by the United Nations as the International Decade for Natural Disaster Reduction, in which earthquake hazard played an important role.

Earthquake safety of a dam is paramount, as it seems that strong earthquakes cannot be predicted in the near future. Water alarm systems at critical dams could save a large number of people from a flood wave but economic and environmental losses cannot be avoided in the case of a dam breach.

The main benefits of a seismic safety evaluation programme for large dams are as follows:

• Compliance with current safety requirements: all parties responsible for the safety of a dam must be insured during very strong earthquakes (legal protection of owners against claims of negligence).

• Socio-economic and political acceptance of dam: the safety of the people living in the downstream area of the dam and their property can be guaranteed.

• Sustainable economic benefits: the dam and the reservoir can be used as initially planned.

Awareness is the first step towards improved earthquake safety. Seminars would be a cost-effective means to raise this and a first step towards a comprehensive programme for the seismic upgrading of existing dams. Technical support for seminars could be provided by a number of international and local dam experts.

Such a programme has, for example, already been implemented successfully in California, US, in the 1990s. During such seminars the optimum methods for the reduction of seismic risk posed by deficient dams can be discussed as well.

The issue needs to be put on the agenda of owners and dam safety agencies. Seismic re-evaluations of older dams, especially those located in seismically active regions, need to be performed and measures taken to improve the safety of any deficient structures.

The dynamic behaviour of dams under the MCE is not known satisfactorily, and so data from strong motion instruments will eventually form the basis for a more reliable seismic safety assessment of existing and future dams.

Today, strong motion instruments and a rapid alarm system should be standard instruments for the safety monitoring of large dams. Complementing the standard static instruments with dynamic strong motion instruments allows for the comprehensive monitoring of a dam under the whole spectrum of actions affecting safety.

In view of the large damage potential of most large dams, it is in the interest of the dam owners, the dam safety agencies and in particular the people affected by a possible dam incident, to reduce the earthquake risk of a dam as far as possible. Predicting the time, location and magnitude of strong earthquakes which may affect the safety of a dam is not an exact science. The importance of issuing warnings to the population living downstream of a dam is, however, a growing issue.

In addition, today’s strong motion instruments have a large dynamic range, ie they can be used for recording small and high amplitude vibrations ranging from a few micro gs to over ten g (acceleration due to gravity: g = 9.81m/sec2 ).

Data collected from strong motion instruments can be used to check and improve the seismic design criteria of the dam, and to locate micro-earthquakes in the vicinity of the dam. The installation of strong motion instruments in all large dams, which are already equipped with pendulums for deflection measurement, is highly recommended. Three instruments would be the absolute minimum for a large dam as it has to be assumed that one or the other instrument may not be working properly at the time of a strong earthquake.

Additional investment

The number of large dams equipped with strong motion devices is steadily increasing. Because of the long return period of strong earthquakes in many parts of the world, the owners lack proper justification for this additional investment, which may be of the order of US$50,000 for a minimum set up. The annual operation and maintenance costs, including the costs for routine data interpretation, should not exceed 20% of the initial investment.

Any new dam with a large damage potential should be equipped with at least a few strong motion instruments, while the monitoring systems of existing dams should eventually be upgraded. Indeed, the damage caused to a considerable number of irrigation dams during the 2001 Bhuj earthquake in India has shown that there is also a need for strong motion instrumentation at smaller dams which are vulnerable to strong ground shaking.

We have to realise that we have not yet solved all seismic problems. Every time there is another strong earthquake, we have to improve seismic standards and regulations.