New Zealand’s first concrete gravity dam was commissioned in 1956 on the Clutha River near Roxburgh. It represented the progress and hope of a new era, bringing electricity to the masses. At the time, large dams were designed in relative isolation to their environments, with little regard given to future impacts, and none whatsoever given to the ultimate challenge of decommissioning. Today, thousands of ageing dams around the world are nearing the end of their economic life cycle, and dam removal and river restoration is becoming an accepted reality.
Although large concrete gravity dams have a theoretical design life of 80-100 years, the actual lifespan of a dam is determined by the rate at which its reservoir fills with sediment. In severely eroding catchments, millions of cubic metres of sediment can be transported annually. The average lifespan of a large dam in China is 45 years.
The Clutha River is New Zealand’s most volatile in terms of volume and flow variation. Historically, it has transported large volumes of sediment, especially from the Shotover catchment. Flooding events can also transport significant sediment loads from numerous tributaries, including the Manuherikia and the Cardrona Rivers. Despite this, the engineers who designed the Roxburgh dam never expected its reservoir to “silt up” so quickly. There have been unconfirmed reports that the dam’s two low level sluice gates were jammed and rendered inoperable within the first 15 years, indicating reservoir-wide sediment transport to the dam wall.
In 1995, ECNZ estimated that 1.5 million cubic metres of sediment annually had been deposited in the reservoir between 1956 and 1992 (when the Clyde dam was commissioned upriver) totalling some 54 million cubic metres. Reports as to the percentage of sediment in the reservoir vary from 40-80%. The area of the reservoir appears to be shrinking accordingly, from 6 square kilometres to 4.5 square kilometres in a recent study.
The first obvious signs of silting up (aggradation) were visible beds of sediment at the Manuherikia confluence and an ever-expanding shingle-bed on the inside of the adjacent bend. Subsequent sediment beds developed along the sides of the reservoir in the Gorge Creek and Shingle Creek areas of the gorge. The reservoir has highly unfavourable impoundment geomorphology, having both a river confluence and a severe natural constriction at its head. Gold-miners knew this narrow section, some 675 metres below Alexandra, as the “Gates of the Gorge.”
Flooding is the expected consequence of reservoir aggradation. Subsequent flooding events inevitably increase in severity, despite remedial works such as “flushing,” sediment removal or the construction of flood defences.
Alexandra is no stranger to floods. Major pre-dam floods occurred in 1863, 1878 (4,650 cumecs) and 1948, while major post-dam floods have occurred in 1987 (2,088 cumecs), 1994 (2,343 cumecs), 1995 (3,212 cumecs) and 1999 (3,760 cumecs). But before the Roxburgh dam the river level was much lower and the threat to the town more remote. The highest volume flood on record in 1878, measuring 4,650 cumecs, did not climb as far into Alexandra as the highest post-dam flood in 1999, measuring 3,760 cumecs. The 1999 flood was the most devastating, despite peaking at only about 80% of the volume of the 1878 flood.
Records clearly indicate that the profile of the riverbed at Alexandra has risen considerably since 1956 when the Roxburgh dam was built. The historic bridge piers, in normal flow, bear mute testimony to the raised river level, now over halfway up the formally exposed lower arches.
Remedial Strategies to "Buy Time"
A sediment monitoring programme was developed after the acquisition of the dam by Contact Energy in 1996, and a formal programme was implemented between Contact Energy and the Crown in 2001. Monitoring has included cross-section surveys, suspended load sampling of inflows and outflows, and particle size analysis of suspended depositions. As a result, it is known that the reservoir traps all the bedload entering it and 80% of the suspended load.
“Flushing” is the first remedial strategy. Flushing involves lowering the reservoir ahead of higher flows in an attempt to move sediment beyond the constricted areas. It is claimed that this has reduced the flood level at Alexandra by 1.7 m compared to 1994 levels. However, bedload re-distribution tends to decrease with each subsequent flushing cycle, and although some sediment is moved into deeper parts of the reservoir nearer the dam, very little of this achieves sufficient suspension to be washed out into the lower river. Flushing is simply a way of “buying time.”
When flushing proves increasingly futile, the next strategy is the physical removal of sediment at pressure points. There are two critical areas near Alexandra: the junction of the Clutha and Manuherikia Rivers, and the constricted area at the “Gates of the Gorge.” Some sediment removal has been attempted at the junction, but it is hardly practicable to remove sediment from the “Gates” area. Large scale sediment removal from steep-sided reservoirs is cost prohibitive because of the need to transport millions of cubic metres of material to a suitable location.
Sediment can also be removed from the bed of the Manuherikia River, particularly in the Galloway area, to ease the flooding issue there and reduce sediment transport into the confluence, but again this is a temporary measure, and such work can be undone by a single flood.
In the face of the sediment threat, it is easy to understand why the Crown and Contact Energy reached an agreement in 2000 to co-fund the acquisition of flood affected properties or easements in the Alexandra area, and to construct flood protection walls to protect the town. The flood walls provide for flood events up to 143.6 metres above sea level. In many ways, flood protection is an admission of defeat, because clearly it is not feasible to keep building up the flood defences as the sediment accumulates.
But the most desperate measure is to raise the operating level of the reservoir (the current range is 1.85m). This would extend the dam’s viable operating years while increasing the risks associated with flooding events. Simply, while the water would move more freely, more sediment would be deposited at the upper end of the reservoir, and major floods could overtop Alexandra’s flood wall more quickly. Correspondingly, it would take less time for floodwaters to overtop the dam crest if the spillways were overwhelmed. The higher water level could also trigger reservoir-wide impacts, not the least of which could be the activation of known landslide areas. In the life cycle of the dam, raising the operating level could be called the “Russian roulette phase.” (Postscript: In November, 2009, Contact Energy was granted consent to raise the operating level .6m, raising the maximum operating level of the reservoir from 132m to 132.6m.)
Given that cost-effective remedial options are limited, the sediment problem seems set to evolve into an issue of legal liability, lost in a quandary of indecision over how long this situation can continue. Contact Energy will claim, naturally, that they are doing all they can to solve the problem. The Crown will also claim, naturally, that they are doing all they can to protect the public.
Truly, heads are buried in the sand (sediment) over this, because everyone knows that the problem has not been solved, that trouble is ahead, and that someone must pay.
Spillway Capacity and Safety
Moreover, the dam itself is now facing a critical safety issue. The Roxburgh dam was not built to accommodate the kind of extreme flooding events that are becoming more frequent and severe with climate change. Lack of spillway capacity is a leading cause of dam failure, and many older dams are now categorized as “high hazard” for this reason. Worldwide, an increasing number of large dams with inadequate spillways, deemed too expensive to upgrade, are being decommissioned.
“Most alarmingly, the world's more than 54,000 existing large dams have not been built to allow for the erratic hydrological patterns that climate change is bringing. In this sense, all dams should now be considered unsafe. More extreme storms and increasingly severe floods will have major implications for dam safety.” – International Rivers
The old spillways of the Roxburgh dam, according to first-hand accounts, barely coped with the 1999 flood. Can the dam and spillways cope with an even greater flood? Damage undoubtedly occurred during the spilling in 1999, but the extent of this remains a matter for speculation. Certainly, significant erosion occurred at the base of the spillways and remedial work was later undertaken.
Subsequently, it was established that the 3 spillway gates were highly susceptible to seismic failure, and so their counterweights were removed and replaced with more reliable hydraulic mechanisms. The 1999 flood may have alerted staff to the urgency of this issue. One rumour suggested that the dam shook so violently that dam workers were evacuated as a precaution. If this is true, one wonders why the communities below the dam were not warned.
When a dam is overwhelmed by inflows, it suffers maximum stress loading above and beyond its design parameters as spillways strain to release an increasing volume of water. Vibration throughout the dam can cause structural failures in the dam crest, abutment interfaces, and the foundation before or during overtopping. Modelling has shown that the relatively thin dam crest is the weakest structural element and the most vulnerable to movement decay. Even more serious is the potential for foundation displacement, risking total failure.
Liability and Who Pays?
A deluge of issues now arise. If the Roxburgh dam experienced an overtopping event, would it survive? If the dam or an abutment breached, who would be liable? Would the dam owners claim that, since the Government built it, they (the taxpayer) should shoulder most of the burden or all of it? What damage would the torrent cause downriver? If the flooding event triggered landslides in the gorges, what further damage and loss of life could occur? What landslide issues exist in the Roxburgh Gorge given the strong history of instability akin to the Cromwell Gorge? Has the fault-line at Coal Creek weakened the dam? Has the historical leakage (1963-65) recurred in the right-facing abutment rock interface?
Time is running out for the Roxburgh dam, and the burden of responsibility rests with the dam owners, Contact Energy, and the dam regulators, the Crown. Decommissioning a large dam is a complex and expensive process that can take decades to complete. The costs involved are significant, and have been estimated at 35-150% in proportion to the cost of dam construction at current values. Typically, there is no provision for these long-term costs when a dam is built. Such provisions would render most large dams uneconomic from the outset.
Forward planning is vital to clearly establish legal liability, dam removal methods, monitoring systems, and a staged timetable to control sediment loading downriver in order to minimise community and abstraction impacts.
River Restoration Opportunity
It has been said that dam decommissioning spells not only the end of a dam, but the renewal of a river. In this sense, dam decommissioning is a restorative and creative process bringing enormous opportunities to local communities. It will take a decade or more to remove most of the sediment from the Roxburgh Gorge using the natural flow of the river and weathering. The geomorphology of the river will change, immediately upriver, and throughout its downriver course which will no longer be starved of shingle and beach sand. The finer suspended sediment will also reach the coast and begin replenishing beaches as it is carried north by ocean currents. Historically, this contained white quartz particulate, which accounts for the loss of the former white sand beaches remembered by local people.
The rediscovery of the Roxburgh Gorge will make national and international news. Gold-mining relics, and the largest rapids in New Zealand, will gradually re-appear, including the Molyneux and Golden Falls, beckoning the white-water fraternity from around the world. Flooded trails along the gorge, cut by gold-miners, will become passable and will be further developed. Alexandra and Roxburgh will reap the benefits of new recreation and tourism opportunities.
A First For New Zealand?
In New Zealand, for consenting purposes, dams are regarded as buildings. As a first step toward decommissioning the Roxburgh dam, a feasibility study is needed to determine the methodologies of dam removal and de-sedimentation – the largest issue. When this is complete, an application will be required under the Resource Management Act to obtain consents to remove the dam and the impounded sediment.
The Roxburgh dam will probably be the first large concrete gravity dam to be decommissioned in New Zealand. The obvious question is, when? In 2007, Contact Energy was granted consents to continue operating its dams on the Clutha River for another 35 years. It seems highly unlikely that the reservoir will remain viable until 2042. The most likely scenario is that another major flood will prompt an investigation into the decommissioning issue, but probably only after further significant damage occurs. A major earthquake could also hasten this process.
Of course, it would be better to prevent the escalation of sedimentation and safety issues before another disastrous flood, or dam failure. Unfortunately, there is nothing forward-looking about the hydropower industry, and the consenting authorities in New Zealand have insufficient expertise and are disinclined to acquire it and act on it.
It’s time to face the fact that decommissioning the Roxburgh dam is inevitable. The costs and impacts will be substantial, so planning should begin sooner rather than later. To delay is to invite greater costs and risks.