Should dams be modified for the probable maximum flood? If benefit-cost ratios, construction accidents and increased occurrence of annual downstream flood losses are taken into consideration, the answer may well be no
Modifying dams for the probable maximum flood (PMF) can result in waste. This statement is the basis of a paper published recently in the Journal of the American Water Resources Association. Wayne J Graham from the US Bureau of Reclamation gives his own interpretation of what should be deemed as an appropriate PMF policy. ‘A modification done in the name of dam safety sounds good on the surface,’ he says, ‘but are resources wisely spent when dams are modified for the PMF?’
Over the past 50 years general dam safety guidelines in the US have stated that dams built upstream of populated areas should not fail during the PMF. Graham adds that current federal guidelines, and even the laws of many states, require that dams must be capable of withstanding the PMF where danger to downstream life and property during the PMF is greater with dam failure, than without dam failure.
Over 24,000 dams in the US are considered to pose a threat to human life or property due to their size and location. Thousands of these dams, classified as significant or high hazard, were not designed using the PMF procedures and consequently would be threatened by such an event. Indeed, calculations suggest that in 1999 most significant and high hazard non federal dams in the US would be overtopped if the latest calculations for PMF at each dam site were used.
Graham puts such data into perspective. Seven US dam failures have caused more than 50 fatalities since 1874, with numbers ranging from unknown to 2209. Only three of the structures failed from overtopping: South Fork in Pennsylvania (1889), Walnut Grove in Arizona (1890), and Canyon Lake in South Dakota (1972).
Over the past 25 years dam failures have caused an average of five deaths per year in the US, with three of these deaths per year caused by dams that have failed from overtopping. By comparison, over 5000 pedestrians die each year in the US due to motor vehicle accidents.
Various individuals have questioned the logic behind modifying existing dams to pass current estimates of the PMF. Terms such as ‘wasteful’ and ‘unwise’ have been used. Others have called for public discussion on the cost-benefits of ‘blindly’ retrofitting existing dams. As the figures in the table opposite show: risks exist and some believe that society must decide what portion of its resources it wants to allocate to reducing deaths associated with such risks.
Spillway design flood criteria are ‘extraordinarily conservative’, Dubler and Grigg claimed in 1996. They spoke of many instances of unjustified expenses, while in 1998 Lave and Balvanyos suggested that retrofitting existing dams to pass the PMF ‘is almost never a good use of funds’.
The main criticism of the PMF design policy is that all too often, cost-benefit ratios of dam modifications are low — Graham states that it is less than 0.01. So ‘fixing all dams to pass the PMF would be economically unsound’, as the cost of bringing all US dams up to current PMF standards would be in the region of US$75B. Estimates even suggest that the cost of saving a life in many dam safety modifications can sometimes exceed US$1B.
In his paper Graham introduces a procedure for evaluating proposed dam modifications. It can be used for proposed modifications to accommodate the full PMF or floods smaller than the PMF. From this a logical rule would be to reject any modifications that cost more than US$5M per life saved, or if the benefit per life lost was less than US$5M.
Graham says it will sometimes be necessary to analyse modifications for their cost-effectiveness in saving lives, and believes that his procedure can form the basis of a new policy.
But new policies are not always readily accepted. Graham acknowledges that dam professionals do not want to be associated with dam failures which cause death and property damage, while legal and personal liability issues make this even more of a contentious issue. But he argues that the PMF policy has only existed for such a long time as it partly reduces the need to make difficult decisions. Some of this ‘hesitation’ to accept the risk of dam failure can be abated by implementing or adopting emergency action plans or warning systems to reduce deaths from large spillway flows or dam failure.
Do larger spillways make safer dams?
Graham acknowledges that modifications to existing dams can reduce economic losses from dam failure, but they can also lead to economic losses due to large spillway flows in non failure events.
It is suggested that an increase in spillway size is not always accompanied by increased safety. Quite often the fact is overlooked that large spillways can actually make downstream flooding worse. Other factors, such as the catchment, reservoir, dam, channel and damage site, not just the spillway, must be considered.
Construction death toll
A final point which Graham addresses in his paper is that dam safety modifications can actually cause more deaths than the dam failure itself. The international-commission-on-large-dams acknowledges that dam building worldwide causes several hundred job site deaths annually. However, Graham believes that workers should not be involved in any activity that claims more lives than it saves. ‘If this line of thinking prevailed not only within dam safety, but also outside dam safety, everyone would benefit,’ he says.
Estimates suggest that if US$75B was spent on correcting hydrologic deficiencies in US dams, about 105 people would die in construction related accidents as a consequence. In comparison, less than three people have died each year from dam failures caused by overtopping, the very thing that the construction workers would be trying to prevent.
Various individuals believe that the PMF policy has outlived its useful purpose. Graham says that it needs to be replaced by a new strategy that considers both the positive and the negative aspects of dam safety modifications.