Is the Clyde Dam Safe?
This question has been vigorously debated since the discovery of the ‘River Channel Fault’ beneath the dam, after which investigations revealed that major geo-technical hazards exist throughout the Cromwell Gorge.
The fault beneath the dam is 12-15kms deep and is connected to the larger Dunstan-Cairnmuir Fault a few kilometres above the dam. This system is part of the Great Alpine Fault. The pre-eminent New Zealand geologist of the 20th century, Harold W. Wellman (1909-1999), defined the Alpine Fault as one of the major transcurrent faults in the world, and one of the most regularly active.
An international review team of geologists was highly critical of the lack of proper pre-dam construction investigation work. It was not at all clear if and when the dam could be made safe. Some experts were adamant that no amount of remedial work, on the dam design and to the landslide zones, could reduce the risks to an acceptable level.
Gerald Lensen (1921-2004), a colleague of Harold Wellman, and one of the leading scientists on active fault research in New Zealand, strongly opposed siting the dam above the fault. Lensen, despite being well-known and highly respected internationally, was regarded by the Government as a nuisance. His brilliant contribution to the New Zealand Geographical Survey, Department of Scientific and Industrial Research (DSIR), ended in 1981, when he resigned in protest (the official line is that he ‘retired’).
Lensen’s view, that concrete dams should not be built on active faults, is now the accepted international norm, espoused by the International Commission on Large Dams (ICOLD).
Controversially, the dam was re-designed in 1982 to incorporate a “slip-joint” intended to accommodate up to 2 metres of lateral movement and 1 metre of vertical movement, in the event of a major earthquake. However, research has revealed that as much as 8-9 metres of lateral movement has occurred on the site in the past and is possible again. Other research points to an imminent "great earthquake."
Associate Professor Jim Davies, Canterbury University, in a talk given at Cromwell, Wanaka and Queenstown, 8-10 October 2007, stated that:
‘The historical patterns of earthquakes and current research on the Alpine Fault indicate that it is likely to rupture very soon. It is 280 years since the last earthquake. The current pressures in the tectonic plates make it probable that the next earthquake will occur in the next 1-20 years.
With an expected magnitude of 8+ this will be considered a "great earthquake" not simply a strong one. The force will result in a horizontal earth shift of up to 8 metres, and a vertical displacement of 4 metres. The effects will be worst in West Otago, diminishing eastward.
The effects will be amplified in South Island mountainous regions and high country where enormous damage can occur to peaks and ridges. Countless landslides can be expected of all sizes. In areas where the magnitude is plus or minus 9, many tens of millions of cubic metres of rock and scree may collapse from slopes.
Damaging aftershocks are likely to continue for several weeks afterwards and the event will have disastrous consequences across many regions. Less intense shaking will continue for months. Liquefaction and widespread ground damage will occur.’
So what would happen to the Clyde dam in such an earthquake?
A report prepared for the ECNZ in 1995, included a seismic analysis of the Clyde dam, stating:
‘The Clyde dam block stresses and accelerations were estimated using linear elastic finite element methods taking account of reservoir and foundation interaction for both the Operating Base Earthquake (OBE) and Maximum Design Earthquake (MDE) loading cases. Concrete stresses were generally less than 1.5MPa for the OBE and showed that cracking was possible in the MDE. These higher MDE stresses were judged acceptable.’
Given that the MDE (Maximum Design Earthquake) is one that would result in no more than 2 metres of lateral movement, and that the expected earthquake could result in 2 to 4 times this amount of lateral movement, it would appear that the “slip-joint” is poor mitigation, at best. Gerald Lensen also insisted that the fault movement would be tensional (would pull apart), and as such the “slip-joint,” which is not designed for tensional movement, would not work effectively.
If the dam survives the next "great earthquake," liquefaction of the loess within the Cromwell Gorge landslide zones could still cause massive deformation, resulting in waves overtopping the dam and reaching areas up the reservoir around Cromwell.
A staggering $936 million was spent stabilising fourteen major landslide zones in the gorge, to hopefully prevent this scenario. Yet most of the gorge is potentially unstable on both sides and in a regional magnitude 8+ earthquake movement must be considered likely in some areas, including some of the known landslide zones.
So what are the likely scenarios for the Clyde dam?
It is expected that a MDE (Maximum Design Earthquake) resulting in no more than 2 metres of lateral and 1 metre of vertical shift, would cause cracking. The “slip-joint” would be ‘spent’ and possibly leaking. The vertical concrete "wedge" on the reservoir side of the "slip-joint," held in place by water pressure, would be pushed against the two deformed sides of the joint, and intense pressure loading could cause further rupturing. If the fault movement is tensional (apart), the two metre wide “slip-joint” would simply open up and a major failure could occur as the two 102m high dam "halves" separate. An MDE could also trigger one or more landslides in the gorge, compounding any dam rupture.
If the next earthquake exceeds the MDE, it is likely there would be multiple failures in the dam and within the landslide zones in the gorge. The Clyde landslide zone directly above the west side of the dam is a particular concern. The consequences of such failures would be catastrophic and many lives could be lost.
But even if the dam performs to its design specification, and a disaster is averted, what happens to a dam that has used up its design capability to withstand an earthquake?
The Clyde dam was built to last 80 years. Regardless of how long it survives, decommissioning will be an expensive exercise involving the removal of the dam, and the restoration of the Cromwell Gorge. Decommissioning and river restoration costs for a large dam are calculated as a proportion of construction costs, and are between 35% and 150%. De-silting the reservoir will be a major cost component, requiring a staged restoration plan, taking years to complete.
What is often overlooked, is that in any large earthquake, the Roxburgh dam would be at greater risk. There was no specific earthquake mitigation incorporated into the dam design, and there was no landslide stabilisation work undertaken in the Roxburgh Gorge, where large landslide zones are also known to exist. The Roxburgh dam had a leakage problem as early as 1963, and it has already sustained a number of seismic cracks. An active fault runs close to the dam at Coal Creek.
Unbelievably, the politicians responsible for the Clyde dam believed that the potentially catastrophic risks were ‘acceptable.’ Officially, there appears to be a ‘head in the sand’ approach to this risk. There has been, and still us, an astonishing and reckless disregard for public safety. Tragically, the most expensive concrete structure in New Zealand is also, considering the evidence, our largest single man-made hazard.
The grim reality is that no one really knows what will happen in the next major earthquake. This fact alone, is an indictment against the dam builders. Meantime, tension continues to increase in the Alpine Fault, and those who would suffer the most – the people of the Clutha River communities, wait …
References:
More...
Martin Wieland, Chairman, ICOLD Committee on Seismic Aspects of Dam Design, Poyry Energy Ltd., Zurich, Switzerland.
A. Bozovic, Former Chairman, ICOLD Committee on Seismic Aspects of Dam Design, Consultant, Belgrade, Serbia.
R.P. Brenner, Past Chairman, ICOLD Committee on Dam Foundations, Consultant, Weinfelden, Switzerland.
‘Natural event and human consequences in Queenstown Lakes and Central Otago’ Tim Davies, Associate Professor, Canterbury University, Mauri McSaveney, GSN Science, 2007.
‘Risk assessment earthquakes, volcanoes, floods and dams in New Zealand’
M.D. Gillon, Electricity Corporation of New Zealand.
‘Seismic Considerations for the Design of the Clyde Dam Transactions’
IPENZ Vol. 14 3/CE, November 1987. Hatton J.W. and Foster P.F.
‘Dams and Earthquakes in New Zealand’
Bulletin of the NZ National Society for Earthquake Engineering, Vol.1, No.2, June 1978. Hatrick A.V.
Chapter 7, Fault Provisioned Design Examples, 7.1 Mitigation Measures, after Bray, 2001, Hamada, 2003.
Hatton, J. W., Black, J. C. and Foster, P. F. (1987). “New Zealand’s Clyde power station,” Water power & Dam Construction, 15–20.
Hatton, J. W., Foster, P. F. and Thomson, R. (1991). “The influence of foundation conditions on the design of Clyde dam, “ 16th Conference on large dams, 157–177.
