Background
Damming rivers has been widely practiced for centuries. Chatterjee (1997) estimates that around the world there are over 40,000 hydroelectric dams that are over 15m tall, and a further 800,000 smaller hydroelectric dams. These numbers do not include dams constructed for other purposes. There are likely hundreds of thousands, if not millions, of dams around the world that exist for the sole purposes of providing water for irrigation, recreation, or flood control. Most large dams are found in developed countries such as the USA, Canada, and Europe. For example, the USA alone contains over 75,000 ‘large’ dams (Graf, 1999; Watson, 1999). Likewise, there are more than 400 dams greater than 1.8m tall and several thousand smaller dams in New Zealand (MfE, 1997). A rapid increase in hydroelectric dam construction exists in developing countries such as Argentina, Brazil, Turkey, Pakistan, and China. China, in particular, has nearly completed construction on the world’s largest hydroelectric dam project, located on the Yangtze River. The World Bank funds many of the large scale hydroelectric projects currently underway. The development of these facilities is one of the important outcomes of their energy policy, which aims to reduce dependence on combustible fuels (World Bank, 2000).
The surge in dam building activity over the last century has largely been the consequence of increasing human populations and by demand for renewable energy sources as people seek to reduce dependence on fossil fuels. Current estimates suggest that hydropower generates approximately 20% of the world’s electrical supply (NHA, 2000). Hydroelectric power is particularly important for countries such as Sweden and New Zealand where it contributes 99.8% (Mellquist, 1985) and c. 79% (MfE, 1997) of the total energy supply respectively. The relative importance of hydropower is comparatively minor in the USA, but this can vary by region. For example, while hydropower provides only 10-12% of the overall energy supply in the USA, it generates 70% of the electricity supply in the Pacific Northwest. Hydropower is also the most significant source of renewable energy in the USA: supplying c. 81% of renewable electricity generated in the USA (NHA, 2000).
Currently, the only other significant alternative to fossil fuel based power production is nuclear power generation. However, social and political concerns regarding safety and long term radiation problems make it difficult to get permission to build reactors. In part this is because methods to solve the long-term implications of highly irradiated sites have yet to be found. Even when permission can be obtained, conditions usually exist that some of the operating profits must be set aside in a dedicated fund. This is to pay for the eventual decommissioning and subsequent cleanup of the reactor site. Interestingly, while nuclear plants are recognized as being temporary structures lasting only a few decades, dams seem to be thought of as permanent fixtures in the landscape. As a consequence, dam owners are seldom obliged to put aside money for the eventual decommissioning and removal of dams. This has important implications as dams age. Contrary to the common perception, dams have finite life spans dictated by the rate of sediment accumulation in reservoirs, and the durability of the building materials used to construct them.
Damming rivers creates far less public concern about safety than does nuclear power. The positive public image of hydroelectric power, combined with the idea of cheap, renewable electricity, means that public opposition to hydroelectric dam construction has generally been muted. Hydropower generally enjoys a public image of being environmentally benign. Of course, the hydropower industry maintains that, for the most part, this image is well deserved. For example, the National Hydropower Association (2000) claims that:
"Hydropower is an emissions-free, renewable and reliable energy source that serves our national environmental and energy policy objectives. With zero air-emissions, hydropower helps in the fight for cleaner air. Hydropower's fuel—water—is essentially infinite and is not depleted in the production of energy."
And also…
"One strategy to reduce U.S. emissions is to develop low carbon or non-carbon electricity generation alternatives. A well known, commercially available source of renewable, emissions-free electricity with extensive development potential is hydropower. Unfortunately, this clean, renewable and reliable energy source is often overlooked and sometimes devalued as a solution to the problem of greenhouse gases. It may be that the magnitude of its contribution to emissions avoidance is too little understood."
However, the perception that dams are environmentally benign is being seriously questioned by some groups such as American Rivers and Trout Unlimited, as well as the oft-maligned fossil-fuel based electricity generators (Bradley, 1997). Environmental organizations concerned about anadromous fish declines have even begun campaigning for the removal of many dams—including some modern hydroelectric dams in the Columbia basin.
The issue of how dams affect rivers and their associated ecosystems, has become increasingly prominent, especially in the last two decades. With an increase in public environmental awareness, and an accompanying toughening of laws, hydroelectric power generators find themselves having to defend their environmental record far more vigorously. Federal Energy Regulatory Commission (FERC) relicensing procedures now mean that dam owners must meet tough new environmental regulations. These include stringent fish-passage measures that were never considered, or provided for, in the early 20th century when many dams were built. Original licenses lasted for a period of 50 years. Relicensing is typically for periods of 30-40 years. The National Hydropower Association reports that new conditions imposed to gain relicensing usually result in an average reduction of 8% of the generating capacity. Approximately 2/3 of non-federal hydropower production (generated by c. 2,500 dams) is due to be re-licensed within the next 15 years (NHA, 2000).
Alongside the issue of relicensing, advancing age presents a concern for many of the dams in America. Approximately 25% of all dams in the USA are greater than 50 years old (Joseph, 1998; American Rivers, 2000). The deterioration of dams over time can lead to serious safety issues. Some 10% of dams are considered ‘high-hazard’ because a dam failure would probably result in loss of human life (Joseph, 1998). Other issues relating to dam safety and potential breaches include extreme storm events that may undermine supporting banks or levees, and seismic activity that may crack and weaken dams, potentially leading to catastrophic failure. Safety issues require that deteriorating or unsafe dams must undergo repairs and maintenance to minimize the risk to public safety. The expense of the work required can actually exceed the value of the dam in some instances, and consequently some dam owners have simply walked away from the problem rather than pay for repairs. These ‘deadbeat dams’ then become the problem of local authorities (or government agencies) who must then determine their fate, and find funding to accomplish it. An example of such a case is Rat Lake Dam in Washington State (American Rivers, 2000). Ownership of this dam was abandoned sometime in the 1960’s. The dam became structurally unsafe, and a potentially affected town was forced to remove the dam as a least-cost measure. The fact that dam removals can be a cheap alternative to repairing or replacing dams appears to have been the motivation behind a recent trend in some states where old, derelict dams have been removed. Wisconsin, in particular, has quietly removed several low-head dams in the last few years. The fact that environmental benefits may accrue from removal, however, has not been lost on the agencies involved and Wisconsin has arguably led the country in evaluating the results of dam removal (e.g., Kanehl & Lyons, 1997).
Apart from the economic factors, some environmental groups argue that relicensing old, inefficient dams is irresponsible because the negative effects on anadromous fishes outweigh the sometimes negligible benefits of such dams (American Rivers, 2000; Watson, 1999; Anonymous, 2000). For example, the Edwards Dam in Maine generated only 0.5% of the State’s electricity requirements and blocked hundreds of kilometers of spawning habitat for several species of fish. In this instance, FERC agreed with the environmental groups and state wildlife agencies, and controversially ordered that the owner must remove the dam.
Many species of anadromous fish in the USA have been listed as threatened or endangered under the Endangered Species Act, and there exists no doubt that some populations of anadromous fish have declined remarkably during the past century (e.g., DOI, 1996; Yoshiyama et al., 1999). In the 1960s, for instance, even after construction of dams on the Columbia River, some 100,000 steelhead and salmon per year migrated up the Columbia's tributary, the Snake River. In 1998, several years after the construction of four dams on the Snake, biologists counted only 9,300 steelhead, 8,426 spring-summer Chinook salmon, 927 fall Chinook, and 2 Sockeye salmon (Kenworthy, 1999). The US federal government now classifies all these species as threatened or endangered, and in the 1980s declared the Snake River coho salmon extinct. Joseph (1998) reports that one hundred and six separate stocks of salmon and trout in the western states are now extinct. Scientists at Oak Ridge National Laboratory (2000) hold federal dams primarily responsible for reducing the Pacific Northwest salmon population from 16 million to 300,000 wild fish per year.
The precipitous decline of fish stocks raises questions about the acceptability of retaining some of the large, and sometimes quite modern, dams that affect them. Contrary to the opinions of many fisheries biologists (e.g., Oak Ridge National Laboratory, 2000), many dam operators maintain that overfishing at sea and oceanic conditions are primarily responsible for the decline in salmonid populations on the west coast. While overfishing and unfavorable oceanic conditions are certainly concerns for anadromous fish populations, the Plan for Analyzing and Testing Hypotheses (PATH) (set up by the Clinton administration to review causes of the decline and provide scientific direction for recovery plans) concurs with the scientists at Oak Ridge National Laboratory. They point to evidence where the decline of runs that are unaffected by dams are far less severe than for other runs in the same river where dam passage is an issue (PATH, 1999; Yoshiyama et al., 1999). If oceanic conditions and over-fishing were the primary cause of the decline then both runs would be similarly impacted. Similarly, Yoshiyama et al. (1999) finds that the least dam-affected run of Californian Chinook Salmon is the only run that is not imperiled.
An Act of Congress (Public Law 102-495) has ordered that two dams on the Elwha River (WA) be removed if it is necessary to fully restore the once bountiful salmon runs in the river. However, political opposition by those who control the federal purse-strings has meant that no federal money has yet been made available to carry out the removal.
Even more controversially, a battle is raging over the fate of four dams on the lower snake River, also in Washington. While PATH, State Departments of Fisheries & Wildlife (WA, ID), Department of the Interior (DOI), and the US Environmental Protection Agency (EPA) all state that removal of the dams is necessary to protect the vanishing runs of salmon (particularly the fall run), the National Marine Fisheries Service (NMFS) and the US Army Corps of Engineers remain opposed to the plan. Their opposition is based on the grounds that the estimated benefits of dam removal are not sufficiently greater than benefits estimated for manually transporting juvenile salmon downstream (NFSC, 1999), to justify the lost hydropower generation. Additionally, the NMFS reinforces their position by highlighting the uncertainties regarding the actual benefits of dam removal for restoring fish populations (NFSC, 1999; NMFS, 1999). This can be partly attributed to the newness of dam removal as an ecological restoration tool. Without previous experience, many of the benefits of dam removal are based on assumptions and predictions.
Recently, the Clinton Administration decided to follow the recommendation of the US Army Corps of Engineers to defer removal plans for 5-10 years to see if lesser measures succeed in restoring the salmon runs (Pope & McLure, 2000; London, 2000). PATH has roundly criticized this recommendation, accusing NMFS of selectively using data to bias their findings and ignoring those that strongly show dam removal as the only option likely to succeed (PATH, 1999). The EPA also criticizes the decision, firstly because of poorly considered assumptions about the thermal effects of the dams (McLure, 2000), and also because the cost estimates provided by the US Army Corps of Engineers to the Clinton Administration failed to include significant costs associated with complying with the Clean Water Act if the dams are not removed. The US Army Corps of Engineers still disputes that it is legally bound to comply with the Clean Water Act (Horsey, 2000) despite an earlier (March 24) ruling in a federal court that they must operate the dams on the lower Snake River in compliance with the Clean Water Act (ENN, 2000).
Meanwhile, environmental groups and many scientists worry that a 5-10 year delay in removing the dams will only lead to more extinctions and a more difficult recovery process. This opinion is echoed in the first two statements of a unanimously accepted resolution of the Oregon Chapter of the American Fisheries Society (AFS, 2000):
1 The four lower Snake River dams are a significant threat to the continued existence of remaining Snake River salmon and steelhead stocks;
2 If society-at-large wishes to restore these salmonids to sustainable, fishable levels, a significant portion of the lower Snake River must be returned to a free-flowing condition by breaching the four lower Snake River dams, and that this action must happen soon"
Clearly the debate regarding removal of large hydropower dams is highly charged and has important social and political, as well as ecological, implications.
One of the major problems facing agencies who must make decisions on whether to remove dams is the lack of good information regarding the actual effects of dam removals, and their alternatives. A lot of planning is currently based on predictions and models, and this reduces the reliability and certainty of decisions based on them. The stance taken by the NMFS in its recommendation to retain the lower Snake River dams exemplifies this problem. Uncertainties regarding transportation-induced, delayed mortality meant that the extreme best case scenario for transporting juveniles downstream was highlighted to minimize the probable difference between dam removal and the alternative barging strategy. This led to media reports downplaying the key finding that dam removal is the only option likely (>50% probability) to succeed in restoring salmon populations in the Snake River. For example, one NMFS press release (NMFS, 1999) acknowledged that :
"Although the scientific study suggests that under the widest set of assumptions, the drawdown or breaching of the four federal Snake River dams may be the most "risk-averse" alternative,…"
but added….
"…the report finds that there are significant uncertainties associated with these projections...Under certain scenarios, there is little or no improvement from dam breaching."
In this case, data regarding transportation mortality rates (which still resulted in a lower chance of success under the extreme best case scenario) was viewed more favorably by the agency than the assumption based benefits predicted for dam removal. The reason for the bias seemingly exhibited by the NMFS is unclear, but it is clear that lack of clear-cut evidence for either proposition allowed sufficient latitude for maneuvering the decision to accommodate interests other than the environment.
This paper attempts to draw together information that is presently available but difficult to access, so that an overview of what is already known about the broad effects of dams and dam removals is available. Hopefully this will identify future research directions that will enable decision makers to base decisions on evidence, and not assumptions. Ideally, published literature alone would provide the basis for such a discussion but the limited availability of scientific articles on the topic of dam removal makes this impossible. Therefore the conclusions of this report must be viewed with some skepticism until relevant published information becomes available. Additionally, some effects of dam removal can be predicted as a matter of deduction. For example, gas supersaturation is a problem associated with spillways on tall hydropower dams. Removing the dam cannot conceivably fail to solve this problem. For this reason a review of the environmental effects of dams is included before addressing the topic of dam removal effects specifically