The Environmental Effects of Removing Dams
Experience has shown that without adequate study and care, dam removal can have severely adverse impacts. The normal process for removing dams begins with the controlled drawdown of the reservoir to very low levels. Once this is complete, the dam can be breached and removed without the threat of a downstream cataclysm. Breaching and dismantling of the dam can be achieved in a number of ways ranging from dynamite to jackhammers, depending on the size and composition of the dam itself. Finally, debris from the dam is removed and the river allowed to flow freely once more.
Depending on their natures, the impacts of dam removal may be felt at different times. Effects that occur during the removal process itself are usually temporary and limited to disturbance of wildlife in and around the demolition site caused by dust, noise, and the presence of large numbers of people. While such impacts are real and must be considered in determining the timing of dam demolition (i.e. should avoid disturbing nesting birds etc), they are the sort of effects that arise from any large scale construction or demolition site and are not specifically addressed in this essay. However, some other impacts discussed here may also start to occur during the initial drawdown of the reservoir or dam breaching, and control measures to mitigate these should be considered before removal begins, and possibly implemented as part of the removal process itself.
As with dam effects, impacts arising from the removal of dams can be complex and strongly interdependent. Unlike the effects of dams, however, there exists a dearth of research on this subject and consequently it is more difficult to come to grips with the complexities of the topic. It is acknowledged that the effects of removing any particular dam can depend on a wide range of factors. For instance, the benefits of removing dams from pre-existing lakes are probably not the same as for dams on previously free-flowing river reaches. This section intends to summarize key findings and prevailing opinions pertaining to the effects of removing dams.
Removing Physical Barriers
Removing dams solves many migration problems for fish and invertebrates within river systems. The absence of hard-to-bypass obstacles such as dams and lakes anadromous and freshwater species should be able to resume their normal migratory patterns without impediment. Studies have shown that anadromous species are quick to exploit newly available habitat. For example, Bryant et al. (1999) showed that fish ladders allowed coho, pink, chum, and sockeye salmon, as well as cutthroat trout, steelhead, and Dolly Varden, to access previously unavailable spawning habitat upstream of a 7m waterfall in Margaret Creek, Alaska. Likewise, there exists good evidence that the removal of dams enables migratory fish species to recolonize areas of their historical range. For example, the removal of the 4.3m high Woolen Mills Dam from the Milwaukee River in Wisconsin, allowed smallmouth bass from below the dam site to take advantage of spawning areas above the former impoundment (Kanehl et al., 1997). By removing the dam, river fragmentation was reduced and overall fish productivity significantly increased. However, the Milwaukee River remains highly fragmented by other dams and this means that it is still inaccessible to anadromous fish species. Other dam removals in Wisconsin have shown dramatic increases in fish diversity and abundance. For example, removal of the Waterworks dam on the Baraboo River increased fish diversity from 11 species to 24 species in the former impoundment. Likewise, abundance of smallmouth bass increased from 3 to 87 fishes in the same reach (Wisconsin State Department of Natural Resources, 2000).
Anadromous species such as striped bass and American shad have historically spawned as far as 435 km up the Neuse River in North Carolina. Research using telemetry and individually tagged fish (Beasley and Hightower, 1998) showed that the low-head Quaker Neck Dam, the most downstream dam on the river, was a significant barrier to striped bass and American shad. Removal of the dam in mid-1998 opened up 127km of previously restricted river habitat to the two species. Radio telemetry studies in the following two spawning seasons revealed that more than 50% of tagged individuals, of both species, migrated past the former dam site (Bowman & Hightower, 2000). While some individuals migrated reasonably long distances above the former dam, average distances traveled past it were not very large (c. 10km) and this suggests that these species may take time to recolonize the full extent of their former range. Also, species such as white shad, hickory shad, alewife and blueback herring are now present and spawning in sites above the former dam site (Wayne Jones, North Carolina Wildlife Resources Commission, pers. com.) Clearly, removal of the Quaker Neck Dam has had significant short-term benefits for upstream fish diversity in the Neuse River.
Following the success of the Quaker Neck Dam removal, other dams on the Neuse River have also been demolished recently (Rains Mill Dam and the Little River Dam). The removals aim to increase the available habitat for anadromous fish, and also to protect endangered invertebrates such as Tar Spiny mussels and Dwarf-Wedge mussels that use these fish as intermediate hosts (Anon, 1999; Mike Wicker, NC Fisheries and Wildlife Service, pers. com.). However, results from these removals cannot be determined at this early date.
Anadromous fish have also benefited from the removal of the Edwards Dam from the Kennebac River in Maine (Anonymous, 2000; Joseph, 1998; American Rivers, 2000). While mostly based on anecdotal evidence, the accounts are compelling. For example, striped bass, alewives, and even sturgeon have been sighted in waters upstream of the former dam site. Alewives and striped bass in particular have returned in high numbers to upstream reaches (where they were previously absent), and are being caught by recreational anglers. Sturgeon have not been present above the dam for over a century, but have returned since dam removal. Also, unpublished data from the Maine Department of Natural Resources that shows that American Shad and Blueback Herring have successfully spawned, and that Atlantic Salmon have been seen upstream from the former dam site (Matthew O’Donnell, Maine Department of Marine Resources, pers.com.) While technical studies have not yet been released to verify these anecdotal accounts, fisheries workers on the river agree that the removal of the dam has meant an upstream increase in the range of these species (Steve Brooke, American Rivers, pers. com.)
Likewise, removal of the Welch Dam from the Cannon River in Minnesota enabled several species including muskellunge, flathead catfish, bowfin, longnose gar, mooneye, and gizzard shad, to recolonize upstream reaches above the dam site (American Rivers Website, 2000). Fisheries scientists acquainted with the removal of Welch Dam confirm that fish species have been captured above the former dam site that were not present before the removal (Tim Schlagenhaft, Luther Aadland, Minnesota Department of Natural Resources, pers.com.) and that up to 20 species now utilize the rapids that were created by the removal of the dam.
The only documented removal of a dam from a natural lake was the removal of Dead Lake Dam from the Chipola River in Florida. This removal resulted in greater fish diversity (61 cf. 34 species), increased abundance of desirable fish species, and allowed the return of striped bass in reaches above the former dam site and impoundment (Hill et al., 1994).
In contrast to the success stories above, the 1970 removal of the 16.5 m Sweasey Dam from the Mad River in California has not resulted in improved returns of salmon, or steelhead, to historical spawning grounds above the dam (Tuttle, 1994). This failure is attributed to a massive sediment mobilization (2.4 million cubic yards over two years) which filled in downstream pools and riffles that provided necessary migration routes. The Sweasey Dam removal holds considerable importance because it is probably the tallest dam for which any information on removal impacts is available. Lessons from this removal have particular relevance for the proposed removals of other high-head dams such as the Glines Canyon Dam on the Elwha River, or the dams on the Lower Snake River, in Washington State. Fisheries biologists are stated to believe that salmon and steelhead runs in the Mad River will ultimately benefit from the removal of the Sweasey Dam despite the lack of success in the 21 years following the removal.
There exists one glaring omission in the scientific literature regarding dam removal effects. To date there are no comparisons between dam-affected rivers, and the same rivers from which dams have been removed with regard to survival rates of larval and juvenile anadromous fishes on their migration to the sea. While the removal of artificial lakes, elimination of gas supersaturation, restored thermal regimes, and increased turbidity are all likely to significantly boost juvenile survival rates, there has been no investigation to quantify the actual increase in juvenile survival. This lack of information creates one of the most important problems facing river managers under pressure to make decisions about the fate of dams.
Most case studies suggest that the removal of dams provides significant benefits for populations of anadromous fish and other fish species that undertake migrations along river systems. As expected, removing dams tends to reduce populations of fish, such as Carp, that are adapted to still water conditions. However, the fishery benefits of dam removal may also depend on whether the river has other dams remaining (particularly downstream), if dam reservoirs were historically lakes or rivers, and also if the method of dam removal is appropriate and well-managed.
Removal of Artificial Lakes from River Systems
Restored Community Structure in former impoundments
Community structure in the river reaches in former impoundments shows a marked trend away from lentic species to lotic species. For example, prior to the removal of the Waterworks Dam from the Baraboo River (WI), carp dominated the fish community in the impoundment, with only one species of darter (a group that prefer flowing water) and a total diversity of only 11 species. The invertebrate fauna consisted of primarily Chironomidae (Bloodworms; 53%), Oligochaetes (Worms; 28%), Corixidae (Waterbugs; 14%), and Ephemeridae (mayflies; 4%). Aside from the mayflies, these fauna primarily represent still-water (lentic) habitats. After dam removal, however, fish diversity increased dramatically to 24 species. This increase in diversity resulted from more species that prefer flowing water environments (e.g., 4 more species of darter). Similarly, even though chironomids (bloodworms) still dominated the invertebrate fauna (49%), more lotic species had colonised the former impoundment (Hydropsychidae (caseless caddisflies) 32%; Heptageniidae (Mayflies) 9%; Tipulidae (fly larvae) 4%; miscellaneous 6%) showing that the community had become more lotic in composition (David Marshall, WI Department of Natural Resources, unpublished data). Offsetting the increase of lotic fish species, carp numbers in the reach were dramatically reduced. A similar pattern occurred following the removal of the Woolen Mills Dam from the Milwaukee River in Wisconsin. Specifically, carp numbers were greatly reduced while the abundance of smallmouth bass greatly increased (Kanehl et al., 1997). Striped bass, largemouth bass, and other lotic species, all increased in abundance after the removal of Dead Lake Dam in Florida, while overall fish diversity skyrocketed from 34 to 61 species (Hill et al, 1994). Likewise, lotic invertebrates such as mayflies increased in abundance (225 cf. 5 per sampler), and stonefly larvae recolonized (50 cf. 0 per sampler) reaches of the Kennebac River in the former impoundment of the Edwards Dam (Maine) within two months of dam removal. The diversity of the invertebrate fauna also increased, from 18 to approximately 60 species, while overall abundance went from 50 individuals per sampler to over 2,000 individuals per sampler (Susan Davies, Maine Department of Environmental Protection, unpublished data). From the data available, it is clear that there is a pattern for increased diversity and abundance of species in former impoundments, and that lentic species are generally supplanted by lotic ones.
Some consider increased numbers of waterfowl, and also raptors such as the peregrine falcon and the bald eagle, to be one of the foremost environmental benefits of dam reservoirs. It is widely assumed that removal of dams will result in localized population reductions of these birds. At present this claim suffers from a lack of documentary evidence, although it appears to be a reasonable deduction given than bird life has been found to dramatically change after construction of dams (e.g.,Stevens et al., 1997). However, this pattern may not always be true. For example, bald eagle abundance has actually increased in the vicinity of the former dam site of the Edwards Dam on the Kennebac River in Maine. Some observers attribute this increase to the presence of previously submerged gravel bars that seagulls use as a refuge from terrestrial predators. Bald eagles have been observed preying upon the seagulls while they rest (American Rivers Website, 2000).
Riparian vegetation and marginal wetlands
The impact of dam removal on riparian vegetation and marginal wetlands has yet to be formally documented but environmental impacts assessments done in the USA generally assume that marginal wetlands and plant diversity will increase over the long term in response to restored flow levels and patterns of variability. Hill et al (1994) comment upon a ‘rejuvenation’ of backwater, wetland areas that occurred after the removal of Dead Lake Dam which suggests that this assumption is a valid one.
Restored conditions in former impoundments
Typically, accumulated sediments in former impoundments are scoured out once free-flowing conditions are restored. Consequently, affected river reaches usually return to their historical conditions. Depending on where in the river system a dam was situated, this may mean a change from silt and sediment to gravel substrates, and the presence of pools and riffles. Examples of such changes include the removal of Edwards Dam from the Kennebac River in Maine. Removal of the dam revealed previously submerged rapids and exposed gravel bars in the former impoundment (American Rivers Website, 2000). Likewise, three-quarters of a mile of riffle habitat was revealed after the removal of the Clyde River Dam in Vermont (Watson, 1999). Bank stability, fish cover, and rocky substrate percentage all improved after removal of the Woolen Mills Dam in Wisconsin (Kanehl et al., 1997). Even for the disastrous removal of the Sweasey Dam from the Mad River, California, habitat conditions in the former impoundment improved (Tuttle, 1994).
Intuitively, the removal of dams from natural lakes is less likely to have a significant impact on habitat quality because still water conditions will persist. The removal of Dead Lake Dam produces the only case where impacts of dam removal from a natural lake have been studied. However, it is unlikely to be a typical example because Dead Lake is really just a very shallow, wide reach of the Chipola River (Hill et al, 1994). Most natural lakes are considerably deeper and less prone to drying out as extensively as Dead Lake. Even so, this example provides a salient reminder that impacts of any dam removal should be considered on an individual basis.
Dead Lake was dammed in 1960 to increase lake storage for drought relief purposes. Initially, the sport fishery improved after dam construction. It slowly declined as aquatic plants choked the lake, and stratification of dissolved oxygen became a problem. It was then discovered that the cyclical drying of exposed lake-bed sediments during periods of low flow was critical for the control of aquatic plant growth, and that this in turn ensured suitable nursery habitat for juvenile fish. Also, exposure of the sediments to air oxidized organic material and reduced biochemical oxygen demand, thus helping to alleviate the demand for dissolved oxygen in the lake. When the dam was removed, fish diversity increased dramatically in the restored impoundment, and the dissolved oxygen levels improved greatly. Also, excessive aquatic plant growth was significantly reduced, and marginal wetlands were restored that provide important nursery habitat for juvenile fishes (Hill et al., 1994).
Downstream releases of stored sediment
The release of stored sediments is probably the most widely recognized problem associated with removing dams. Shuman (1995) reviewed sediment mobilization processes, based on several dam removal case studies, to provide direction for other workers undertaking environmental impact assessments for dam removals. The case studies examined demonstrate that the impacts of sediment can be severe unless appropriate mitigation measures are put in place.
When vast quantities of accumulated sediment are suddenly released, it can have adverse effects such as increased turbidity, smothering of downstream substrates, release of toxic substances, and deoxygenation. These effects can be short term and transitory, but they may persist for decades if removals fail to mitigate or prevent massive sediment releases.
Kanehl et al. (1997) reported short term problems resulting from sediment released during the removal of Woolen Mills Dam from the Milwaukee River. While the sediment release was compounded by a simultaneous diversion of the downstream watercourse, significant reductions in the downstream populations of fish were observed following dam removal. However, the affected fish populations reportedly recovered within two years of the removal, thanks to extensive bank stabilization works in the former impoundment that prevented ongoing releases of sediment because of channel erosion.
This problem is not only important for dam removal projects, but also for projects that try to desilt or dredge reservoirs that have become filled with sediment. For example, attempts to desilt a small weir in south Australia resulted in extremely high turbidities downstream, thick deposits of sand extending below the weir, and finer fractions depositing as far as 2 km downstream (Doeg & Koehn, 1994). The invertebrate fauna, 2.1 km below the weir, suffered an immediate 20% reduction in diversity, and 64% reduction in abundance. Rainfall events that re-mobilized the sediment, and also further desilting works, set back invertebrate recovery in the affected reaches. Fish populations in the stream also suffered, with a 93% reduction close to the weir and 51% fewer fish 2.7 km downstream. However, a very large storm event eventually scoured the weir reservoir clear of sediment and moved much of the deposited sediments further downstream. Recovery of the invertebrate fauna after this event was relatively swift, but fish populations were reported to be recovering much more slowly two years after the storm flushed out the sediments.
Where planning and forethought are inadequate, the release of large amounts of sediment can have effects that persist for many decades. Two dam removals, in particular, illustrate this point.
As has been previously mentioned, the Sweasey Dam was removed from the Mad River in California during 1970 by blowing the structure up with dynamite after the reservoir water was drawn down (Tuttle, 1994). Its reservoir filled with sediment, the dam no longer served any useful purpose. Dam removal appeared to be more cost effective than dredging the lake free of sediment. No attempts were made to ameliorate the release of the massive amount of sediment stored in the reservoir and, consequently, an immense sediment wave began to slowly migrate down the river. Rainfall events continued to re-mobilize sediments by causing the river to erode the newly formed river banks in the former impoundment, cutting deep into the thick layers of accumulated sediment stored there. Tuttle (1994) estimates that 1.8 million m3 of sediment were released over the first two years. As a result, thick layers of sediment filled downstream pools and the river became much shallower and wider. Without the pool-riffle habitat necessary for negotiating their way upstream, salmon and steelhead could not bypass the long stretches of silted-up river bed. There has been no significant improvement in the runs of anadromous species in the Mad River. According to Tuttle (1994), studies show that the toe of the sediment wave only began to reach the ocean during the mid 1990’s and it may be years before the river finally clears itself of the last of it. However, fisheries biologists are stated to believe that salmon and steelhead will ultimately benefit from the dam removal once the sediment has been cleared and pool-riffle habitat is restored.
Likewise, removal of the 9.1m high Fort Edwards Dam in 1973 from the Hudson River in New York State caused a mass mobilization of sediment (estimated at 336,000 m3 during the first year) and debris that resulted in navigation barriers for naval shipping in parts of the Hudson River (Shuman, 1995; American Rivers Website, 2000). Although the ecological impacts of this mass mobilization have not been reported, it is significant that sediments from this dam contained high levels of polychlorinated biphenols (PCB’s) which are highly toxic, and it is likely that the smothering of downstream reaches, together with the toxic impacts from the PCB’s, had significant adverse effects. The pre-removal assessment failed to adequately address the nature of the sediments because only sediment from the upper layers of the reservoir bottom were analysed. PCB discharges into the river had ceased prior to the deposition of the surface sediment layers and consequently were not discovered until too late. PCB transport from the reservoir is still monitored to this day.
Overall, post-removal sediment waves (Shuman, 1995) and ongoing releases of impounded sediment are of critical concern for the ecological health of rivers in the short-term, but unlikely to be long-term concerns because the river will eventually discharge the accumulated sediments into the sea. However, it is possible that short term impacts could result in the extinction of some species, particularly if the population is already dangerously low and if mitigating measures are inadequate.
Effects on coastal areas
Since dams appear to cause erosion of estuaries and deltas, it seems likely that these areas will start to agrade once sediment transportation processes are restored. This agradation may be particularly rapid after dam removal because of the sudden release of large quantities of previously stored sediment. One example of this is the migration of the mouth of the Mad River in California. Tuttle (1994) reports that the river mouth has migrated nearly 3 miles northwards since the removal of the Sweasey Dam in 1970. However, this is an area which has received little attention in the literature thus far.
Thermal Effects
There exists considerable debate over how removing dams from regulated rivers will actually impact water temperatures since the effect of dams on this parameter can vary. However, it is generally assumed that river temperatures will fluctuate over a wider range and be more responsive to ambient temperatures and conditions. This can lead to contradictory claims about how this will affect instream biota. For example, the NMFS claims that removing the dams on the lower Snake River in Washington State will make the water too warm for the returning salmon (McLure, 2000), but this claim neglects that the timing of salmon runs is dependent on the sudden drop in water temperature that happens in unregulated rivers. Regulated rivers may be cooler overall, but are not as cold in winter and spring when large runs of salmon are migrating upstream. It seems strange that returning rivers to their original unregulated condition could be harmful to fishes that have been living in those conditions for hundreds of thousand of years. The NMFS, however, justify their position by pointing to changes in riparian cover and land-use that may increase the water temperatures from historical levels.
Unfortunately, none of the dam removal case studies examined here compare downstream temperatures before and after dam removal.
Greenhouse gases and water losses
By removing dams and their reservoirs it is expected that anaerobic decomposition of organic matter will be reduced. This should help reduce the production of methane and carbon dioxide but has not been studied.
The extra evaporation associated with large lake surface areas and warm surface water is likely to be eliminated. Likewise, extra infiltration of river water through the large area of flooded terrain is eliminated when dams are removed.
Health Risks
No work has been done to quantify changes to public health risk, or rates of disease for wildlife, resulting from dam removals. While some diseases flourish in lakes, others prefer wetlands and by eliminating one habitat and restoring the other there may be a tradeoff that limits changes in the risk to public health. However, diseases that affect instream biota are more closely associated with lacustrine environments and, consequently, dam removal should reduce the risk of disease to fish.
Removing dams eliminates the risk of catastrophic breaches that could threaten people and wildlife downstream.
Restored River Flows
The removal of dams will clearly restore normal patterns of downstream flow variation and it is widely assumed that this will result in improved conditions for instream biota and marginal ecosystems. However, research to determine recovery rates, and processes, is presently unavailable.
If flood control was one function of a dam then developments may have been built on the old floodplain while the dam was in place. If so, these developments may be at risk of flooding after the dam is removed.
Summary of Dam Removal Effects
One of the foremost reasons put forward in favor of removing dams is restoring the migratory patterns of anadromous fishes. Case studies reveal that the removal of dams does successfully restore longitudinal linkages in river systems, and that anadromous species typically recolonize upstream reaches relatively swiftly. However, the magnitude and success of recolonization depends on a number of issues including the presence of other dams downstream, and the effects associated with sediment release from dam reservoirs. No studies have yet demonstrated that the upstream recolonization by anadromous species successfully increases the total productivity of the river. This is an area that should be addressed to demonstrate benefits to fisheries, rather than simply assuming them. However, one case study (Kanehl et al., 1997) shows that the productivity of freshwater species such as smallmouth bass is enhanced by dam removal, and this bodes well for anadromous species also.
Case studies of dam removals reveal marked changes in community structure in formerly impounded river reaches. Typically, this involves the reduction of species adapted to still water conditions such as carp, pollution-tolerant macroinvertebrates and some aquatic plants. It is generally assumed that waterfowl and raptors also will become less common, although one case study (Edawrds Dam Removal) found that bald eagle abundance in the formerly impounded reach actually increased (American Rivers, 2000). Offsetting these losses, fish and invertebrate diversity in formerly impounded reaches has been shown to dramatically increase, and this increase has come about because of recolonization by species that prefer clean, flowing water. This probably results from changes in the nature of instream habitat such as increases in rocky substrates, fish cover, and the formation of pools and riffles.
The sudden release of sediment stored in dam reservoirs is a matter of particular importance. Previous dam removals have shown that when sediment release is carefully controlled and mitigated the adverse ecological impacts are transitory and persist for only a short time. However, if sediment releases occur on a massive scale then the damage to the ecology of a river may be even more severe than that resulting from the continued presence of a dam, and can persist for decades. Long-term studies on rivers affected in this way are not yet possible due to the relatively recent advent of these dam removals. However, it is widely assumed that rivers will repair the damage by eventually depositing the offending sediment into estuaries and river deltas. Consequently, this is thought to result in either the prograding of coastal areas near river mouths, or at least a reduction in coastal erosion.