Getting Old: Dam Aging and Decommissioning
by Patrick McCully, Zed Books, London, 1996
Once a dam has proved itself well enough built to hold back a reservoir (many dam failures occur either during construction or during or shortly after reservoir filling) its structure and component parts will begin to age. The unique nature of each dam means that every structure will age at a different rate in a different way. Some dams may remain safe for a thousand years, others may start to crack and leak after less than a decade. Around the world, some 5,000 large dams are now more than 50 years old, and the number and size of the dams reaching their half century is rapidly increasing. The average age of dams in the US is now around 40 years. According to a panel on dam ageing at ICOLD’s 1991 Congress, "in the future attention and activity [will] be more and more shifted from the design and construction of new dams to the restoration of the structural and operational safety of existing dams".
The risks to human life caused by the crumbling dams of the ex–Soviet Union are equivalent to the better known risks from its decrepit nuclear power plants. When three engineers from Hydro–Quebec inspected Inguri Dam in the republic of Georgia in 1994, they found the world’s third highest dam "in a rare state of dilapidation". Only two out of five groups of turbines were in working order, the turbine galleries were flooded by water leaking through the concrete arch dam, and the spillway was found to be "defective". Not only does Georgia not have the money for the urgently needed repairs, but at the time of Hydro–Quebec’s visit the staff working at the dam had not even been paid for six months. Inguri was completed only in the 1980s: the condition of many of the older Soviet dams is presumably even worse.
International Water Power & Dam Constructionreported in February, 1996, that water levels behind the Kakhovskaya Dam in the Ukraine had recently receded after reaching "excessively high" levels. Kakhovskaya holds back the world’s largest capacity reservoir. Engineers reportedly warned the Ukrainian government that the failure of the earth embankment would "create a 20–30m high wall of water, moving at a speed of around 130 kilometres per hour" threatening the lives of half a million people living in several cities and hundreds of towns along the Dnieper River.
Dangerous dams, however, are far from purely a problem of the ex–Soviet republics. Between 1977 and 1982 the Corps of Engineers inspected 8,800 non–federal dams in the US, most of them privately–owned, which it classified as "high–hazard" – where a failure could cause significant loss of life. One–third of these dams – 2,900 – were considered to be "unsafe", primarily because of inadequate spillway capacity. A 1994 survey showed at least 1,800 non–federal dams were still unsafe. The situation is similar for federal dams: in 1987 one–fifth of BuRec’s 275 dams were classified as unsafe, as were one–third of the 554 dams operated by the Corps of Engineers.
An Ontario Hydro study of data from several hundred North American dams shows that on average hydrodam operating costs rise dramatically after around 25–35 years of operation due to the increasing need for repairs. When the cost of maintaining an old dam exceeds the receipts from power sales, its owners must decide either to invest in rehabilitating the dam, or, if the cost of repairs would be prohibitive, to disconnect the dam from the grid and cease producing power.
Many old dams in the US have simply been abandoned by their owners. According to the Michigan Department of Natural Resources (MDNR) several abandoned small dams have been washed out during storms in recent years. "These failures," says the MDNR, "have caused extreme erosion, excessive sediment deposition and destruction of aquatic habitat accompanied by the loss of the fisheries". Michigan taxpayers, through the MDNR, have had to pay for removing several "retired" hydroelectric projects, while their ex–owners have suffered no financial liabilities.
Dam decommissioning (defined as anything from merely stopping electricity generation to the expensive and challenging operation of totally removing a dam and restoring the river to its pre–dam state) has recently been forced onto the agenda of an unwilling hydropower industry in the US. More than 500 of the 50–year licenses given by the US Federal Energy Regulatory Commission (FERC) to private hydrodam operators expire between 1989 and 2004. A coalition of river conservation groups have used this spate of expiring licenses to urge FERC to institute a comprehensive dam decommissioning policy. The Hydropower Reform Coalition believes that new licences should only be given on the condition that owners pay into special decommissioning funds during the lifetime of their projects, just as nuclear power plant operators in the US have to put money aside to pay their inevitable decommissioning costs. Despite strong opposition from the hydropower lobby, FERC announced at the end of 1994 that it has authority to order owners of the more than 1800 dams under its jurisdiction to decommission dams which fail to win new licences, although it has not yet conceded the coalition’s call for it to require payments into decommissioning funds.
Although hundreds of dams have been intentionally breached or torn down in the US, the vast majority have been only a few metres high: the largest to be removed is probably the 19–meter Grangeville Dam on the Clearwater River in Idaho, which was dynamited in 1963 to restore salmon runs. How exactly to dismantle a very large dam, what to do with the sediment clogging the reservoir behind it, and how much such an operation would cost, is largely unknown. The Hydropower Reform Coalition believes that removing a hydrodam could cost more than building one. Furthermore, the cost would skyrocket where reservoir sediments contain heavy metals and other toxic contaminants.
The best–known dam decommissioning controversy surrounds a pair of dams on the Elwha River in Washington State: the 30–metre Elwha and 70–metre Glines Canyon Dams. Built in the 1910s and 1920s with a combined installed capacity of 19 megawatts, the dams all but wiped out the once–rich steelhead trout and salmon fisheries of the Elwha, fisheries to which the Elwha S’Klallam Tribe had been guaranteed rights "in perpetuity" in the remarkably aptly named 1855 Treaty of Point No Point. Since the Glines Canyon Dam FERC license came up for renewal in the late 1970s, the Lower Elwha S’Klallam and environmentalists have been trying to get the dams removed. In 1992 their long campaign started to bear fruit when Congress directed the Interior Department to detail the best plan for "full restoration of the Elwha River ecosystem and the native anadromous fisheries". The Interior Department concluded that only removing the dams could fully restore the ecosystem.
Removing both dams and dealing with the 11.5 million cubic metres of sediment which has built up behind them is estimated to cost between $67 million and $80 million, although the full cost of restoring the river, including compensating the dam owners and the paper mill which uses their power, would be between $148 and $203 million over a 20 year period. The dams would be taken down after the river had been diverted around them. Removing the sediment would be the biggest problem and is planned to be done with a combination of dredging, allowing the newly free–flowing river to wash the sediments downstream, and stabilizing with vegetation the sediments higher up the river banks.
No one knows how the major dams built in the last half century will be removed or where the money to pay for this will come from. Feasibility studies rarely – if ever – mention what will happen when a reservoir is choked with sediment or when a company no longer finds it economic to maintain a dam. The international industry has largely steered clear of the issue. "I have only once seen a paper on the decommissioning of dams," Wolfgang Pircher, then President of ICOLD, said in a lecture to the British Dam Society in 1992. It is, however, an issue which sooner, rather than later, will have to be addressed.