Current Threats

Current threats to Atlantic salmon include both the primary and secondary factors identified in the final rule (74 FR 29344). In addition, this recovery plan identifies undersized or poorly designed and installed road stream crossings as a primary threat and – to account for rapidly changing environmental conditions – places greater emphasis on addressing threats to the co-evolved diadromous fish community and to fully functioning and connected habitats.

Factor A. The present or threatened destruction, modification, or curtailment of its habitat or range

As a precursor to assessing habitat-related needs and threats, a standard unit of measure for freshwater habitats was established.  The Habitat Unit (HU) is a common way to describe 100 square meters (m2) of salmon rearing habitat (74 FR 29300). The number of HUs in each SHRU was determined using a GIS-based model.  Based on this model, each SHRU contains the following number of rearing HUs: Penobscot, 315,574 HUs; Merrymeeting Bay, 339,182 HUs; Downeast, 59,066 HUs.  

Lack of accessibility due to dams and road stream crossings, together with mortality from and ecological effects of dams, constitute the major freshwater threat to the recovery of the GOM DPS of Atlantic salmon.  In addition, several secondary stressors combine to threaten habitat needed by DPS salmon.  Both significant and secondary threats are discussed below.

Habitat Quality (secondary threat) 

Habitat quality was determined for watersheds between 40,000 and 250,000 acres in size (i.e., HUC 10) in each SHRU based on key parameters needed for salmon to carry out the freshwater portion of their life cycle.  The NOAA developed a comprehensive habitat quality scoring system based on water temperature, biological communities, water quality, and substrate type as part of the process for identifying critical habitat for Atlantic salmon.  Habitat quality scores are defined as 0 for not suitable; 1 for marginally suitable; 2 for suitable; and 3 for highly suitable.  Only watersheds with scores of 2 or 3 provide critical habitat needed for salmon to carry out their life cycle.  Each SHRU has a significant amount of habitat that was rated suitable for Atlantic salmon spawning with 70 percent, 65 percent, and 60 percent of watersheds scoring 2 or greater in the Merrymeeting Bay, Downeast, and Penobsoct, respectively.  Using the NOAA 2009 data, habitat quality is a lesser stressor on the GOM DPS.   It should be noted that the habitat quality map in figure 1 is intended to provide broad spatial guidance on the location of suitable habitats within the GOM DPS.  These scores reflect the predominant quality of habitat within each watershed and do not reflect whether a watershed is accessible to salmon.

Figure 1. Habitat quality score of HUC 10 watersheds in the GOM DPS (From NOAA 2009). 

Lack of Accessibility from Dams and Road Stream Crossings (significant threat)

 Defining accessible: Access to suitable habitat can be partially or completely blocked by artificial barriers.  For the purposes of this recovery plan, upstream habitat has been categorized for accessibility as follows: 

 Habitat with No Access: Habitat above a barrier (dam or road stream crossing) that has no fish passage

 Habitat with Impeded Access: Habitat above a barrier that temporarily blocks or impairs a salmon’s natural ability to pass (e.g. a culvert or dam with a fishway with limited function).

Habitat that is Accessible:  At a minimum, the habitat must allow for movements of parr that seek out suitable habitats for feeding and sheltering, downstream movements of smolts during the spring migration, and upstream and downstream movement of adults that seek out habitats for spawning and resting.  To meet this standard, habitat must be either:  

 1.   Accessible above a dam with upstream and downstream passage that does not preclude recovery.  This pertains to any project that meets performance standards for fish passage, dam operations, etc. under either section 7 or section 10 of the ESA, thereby indicating that the level of authorized take of DPS salmon is not exceeded.


  1. 2.  Accessible above road stream crossings (e.g., culverts) set at the correct elevation using Stream Simulation methodology (  This approach creates crossing structures that are as similar as possible to natural channel dimensions, slope, and streambed structure; when these characteristics are retained, water velocities and depths also will be retained.  As such, the simulated channel should allow full passage of aquatic animals.


Habitat that is Fully Accessible:  Habitat where there are no artificial barriers between it and the ocean.

Barriers to Fish Passage

Manmade structures (e.g. dams and culverts) fragment habitats and prevent access to upstream areas.  Fragmentation exposes populations that remain in these patches to higher rates of extinction from demographic, environmental, and genetic stochasticity (Lande 1993; Hanski 2002, Drechler and Wissel 1998).  Removing barriers to improve passage at remaining structures significantly decreases fragmentation by increasing access to habitat essential for Atlantic salmon spawning and juvenile rearing.  Increasing access to diverse habitats also restores ecological complexity, allowing salmon to select for more diverse habitat options, which in turn helps protect against environmental stochasticity and helps maintain genetic diversity.

Road stream crossings as an Emerging Significant Threat

Road stream crossings are found throughout the GOM DPS; these structures create passage problems; most frequently in 1st and 2nd order streams.  Corrugated metal, plastic or cement culverts, rather than bridges or bottomless arch culverts, are frequently installed at road crossings to reduce costs.  According to Gibson et al. (2005), culverts create more passage barriers to fish passage than do other road stream crossings (bridges).  Undersized culverts create hydraulic barriers that sever habitat connectivity within the range of the GOM DPS.  Improperly placed and undersized culverts create fish passage barriers through perched outlets, increased water velocities, or insufficient water flow and depth within the culvert.  Poorly placed or designed road stream crossings reduce access to habitat necessary for Atlantic salmon spawning and rearing and alter stream processes including transport of sediment and materials.

Road stream crossings that restrict access to suitable habitat have a significant impact on parr production due to the location of rearing habitat in river systems.  In a study on the Sheepscot River, Sweka et al. (2007) found the percentage of survival from fry to the age 0 parr stage decreased along the Sheepscot River as cumulative drainage area increased and stream reaches of smaller cumulative drainage area contributed more individuals to the total outmigrating smolt population.  Sweka and Mackey (2010) found a similar relationship throughout Maine Atlantic salmon rivers in which parr density decreased with increasing cumulative drainage area, and they modeled this relationship using quantile regression to illustrate an upper bound  (90th percentile) of parr density that could be expected in a stream reach of a given cumulative drainage area. 

Contrasting the spatial distribution of Atlantic salmon habitat and parr productivity in the GOM DPS highlights the importance of smaller streams (e.g. 1st and 2nd order streams).  The distribution of parr production by cumulative drainage area was assessed using a GIS-based habitat model that predicts the amount of Atlantic salmon rearing habitat in stream reaches (Wright et al. 2008) in combination with Sweka and Mackey’s (2010) model that predicts parr density based on cumulative drainage area.  The Atlantic salmon rearing habitat model was developed using data from surveys conducted in the Machias, Sheepscot, Dennys, Sandy, Piscataquis, Mattawmkeag, and Soudabscook Rivers.  The model uses reach slope derived from contour and digital elevation model datasets, cumulative drainage area, and physiographic province to predict the total amount of rearing habitat within a reach.  The GIS model provides estimates of habitat rearing units and drainage area for individual stream reaches.  Figure 10 shows that suitable rearing habitat is distributed fairly evenly across watersheds. 

A different picture emerges when looking at the spatial distribution of juvenile productivity in these drainages.  Atlantic salmon parr production values were calculated by multiplying predicted habitat rearing units by parr density estimates based on cumulative drainage areas using parameter estimates from Sweka and Mackey’s (2010) 90th quantile regression model.  The largest cumulative drainage area used in Sweka and Mackey’s (2010) model was 623 km2 and, lacking estimates for larger drainages, that value was applied to any reaches in the GOM DPS greater than or equal to that cumulative drainage area.  Figure 11 shows that approximately 83 percent of predicted parr production in the GOM DPS occurs in stream reaches with drainage areas less than 100 km2  (90 percent from habitat with drainage areas less than 209 km2, and 95 percent from habitat with drainage areas less than 411 km2). Analysis using stream order calculations derived from ArcGIS using the RivEx tool shows that over 92 percent of the perennial streams in the National Hydrography Dataset in the GOM DPS with drainage areas less than 100 km2 are either 1st or 2d order streams (J Wright, pers. comm.).

The magnitude of the impact of road stream crossings on highly productive habitat becomes apparent when looking at drainages where comprehensive stream crossing assessments have been conducted.    Stream crossing surveys have been completed (Abbott 2008) throughout the Penobscot SHRU except for portions of the West Branch Penobscot River watershed and much of the Merrymeeting Bay SHRU.  Over 5000 stream crossings have been identified in the GOMDPS. Surveys in the Downeast SHRU have not been completed.

1 Information not available (NA)

Add a recovery action to address this to account for changes in numbers of units accessible….numbers provide only an estimate of what is out there. 

Back to Top

Mortality and Ecological Effects of Dams (significant threat)

Salmon smolts and postspawning adults travel downstream to the ocean from spawning and rearing habitat and encounter the same dams their parents experienced on upstream migration to spawning habitats.  Atlantic salmon face routes that pass through, over, or around the dam and facilities.  For electricity-generating dams, salmon can travel over the spillway of the dam, through a downstream fish passage facility if the dam has one, or through power-generating turbines.  Mortality can occur no matter which path Atlantic salmon take, but higher mortality rates occur if fish pass through the turbines.

Direct, indirect, or delayed mortality from entrainment into dam turbines, passage via spillways, or fish bypass has been shown to be an important factor affecting Atlantic salmon.  Dams can also cause indirect mortality.  Smolts injured or disoriented by dams may become more vulnerable to predators after passage (Mesa 1994).  Descaling can also impair osmoregulation, possibly leading to estuarine mortality (Zydlewski et al. 2010).  Lack of flow cues at dam reservoirs can increase the time in the impoundment and thus delay migration and increase predation (Holbrook et al. 2011).

Measuring indirect mortality from dams is difficult, but recent studies are approaching the problem. Holbrook et al. (2011) estimated the survival of outmigrating Atlantic salmon smolts through a large reach of the Penobscot River (up to Weldon dam) in 2005 and 2006.  This study provides a snapshot of total survival through the river, including all causes of mortality (e.g., direct and indirect mortality from dams, predation, disease, fishing, and stress from handling during the research and assessment).  An extensive acoustic array in the river, coupled with release of smolts with sonic transmitters, allowed researchers to separate out survival in specific reaches, including those with dams and those without dams.  They observed high mortality rates in the reaches where fish were released, probably as a result of stress from the surgical insertion of the sonic transmitters.  In addition, high rates of mortality were observed in three other reaches:  Howland dam reach, West Enfield dam reach, and Milford dam reach.  The estimated survival of tagged hatchery smolts through these three reaches with dams was as low as 52 percent, whereas survival in reaches without dams exceeded 95 percent.  Although these data do not definitively reveal sources of mortality, these losses are likely attributable to the direct and indirect effects of the dams (e.g., physical injury, predation).  Shepard (1991) found that for the three lower Penobscot River dams (Milford, Great Works, Veazie) and the intervening habitat, net smolt survival varied between 44 and 94 percent.  For the lower four dams (West Enfield, Milford, Great Works, and Veazie, or for those fish choosing the Stillwater Branch route, (i.e., West Enfield, Stillwater, Orono, and Veazie) the survival rate was between 38 percent and 92 percent.  Given that many Penobscot salmon must pass through at least three dams and sometimes as many as seven, mortality from dams in any single year is likely to be significant.

In addition to the complete or partial blockage to upstream and downstream passage and direct, indirect or delayed mortality associated with passage through dams, dams have a number of additional negative ecological effects on Atlantic salmon.  Dams create impoundments that inundate the natural stream and river habitat and cause sediment deposition that can cover important rearing and spawning habitat.  Impoundments create large pools of water in which water temperatures can increase above preferred Atlantic salmon temperature levels.  These impoundments and associated habitat changes can become preferred habitat for warmwater exotic species that prey on juvenile Atlantic salmon.  Impoundments can cause migratory delays, which, in turn, can cause poor synchrony of physiological tolerance to salinity (McCormick et al. 2009), thereby increasing estuarine mortality (Ferguson et al. 2006).  These effects may be exacerbated by climate change that may also alter predator/prey assemblages by decreasing qualitative habitat features that benefit salmon while concurrently increasing habitat features that benefit predators and competitors.  For additional information, see Fay et al. (2006), and appendix 8 in Fay et al. (2006), and the 2009 GOM DPS Atlantic salmon listing rule (74 FR 29344).

In conclusion, the direct, indirect, and delayed downstream mortality associated with dams and the ecological effects of dams are a significant threat to the recovery of the GOM DPS of Atlantic salmon.

Back to Top

Habitat Complexity (secondary threat)

Some forest, agricultural, and other land use practices have reduced habitat complexity in the range of GOM DPS Atlantic salmon.  Historic timber harvest practices have reduced the abundance and diversity of large wood (LW) and large boulders from many rivers.  Large wood is important for Atlantic salmon during several life history stages.  Nislow et al. (1999) found that survival of salmon fry in small streams in Vermont was strongly correlated with the availability of low-velocity microhabitats, and that the addition of LW (length of LW greater than bank full width) increases the availability of these microhabitats. Large wood may be even more important for older juveniles, because they use stream cover, including LW, particularly during winter (Cunjak et al. 1998).  In winter, salmon may require habitat that provides adequate shelter from adverse physical conditions and protection from predators.  Thus, availability of high quality winter habitat may influence salmon survival during this critical life stage, LW may increase overwinter survival in Pacific salmon by increasing habitat complexity (Quinn and Peterson 1996, Solazzi et al. 2000), and the same mechanism may apply for Atlantic salmon.

Water Quantity (secondary threat)

Direct water withdrawals and groundwater withdrawals for crop irrigation and commercial, and public use can directly impact Atlantic salmon habitat by depleting streamflow (MASTF 1997, Dudley and Stewart 2006, Fay et al. 2006).  Subsequently, reduced stream flow can reduce the quantity of habitat, increase water temperature and reduce dissolved oxygen.  The cumulative effects of individual water withdrawal impacts on Maine rivers is poorly understood; however, it is known that adequate water supply and quality is essential to all life stages and life history behaviors of Atlantic salmon, including adult migration, spawning, fry emergence, and smolt emigration (Fay et al. 2006).

Water Quality (secondary threat)

Maine’s water quality classification system provides for different water quality standards for different classes of water; waters with the best water quality (Class AA waters) have the highest water quality standards, and waters with poorer water quality (Class A, B, and C waters) have progressively lower water quality standards).  These standards were not developed specifically for Atlantic salmon, and the lower quality standards of Class A, B, and C waters may not provide high-enough water quality to protect all life stages of Atlantic salmon.  Some portions of the GOM DPS are in areas with the highest water quality standard (Class AA waters), but many Atlantic salmon are found in areas that are Class A, B, or C, and therefore may be affected by water quality not well suited for various aspects of their life history.

Atlantic salmon may also be impacted by degraded water quality caused by point and nonpoint source discharges.  Point sources are regulated by the DEP, which issues permits under the National Pollutant Discharge Elimination System program of the Clean Water Act that has been delegated by the EPA.  Permits allow for a zone of initial dilution where water quality requirements are reduced.  Salmon in or passing through such zones would be exposed to discharges below water quality standards.

Back to Top

Factor B.  Overutilization for commercial, recreational, scientific, or educational purposes 

As summarized in the 2009 listing rule, overutilization for recreational and commercial purposes was a lesser stressor that contributed to the historical declines of the GOM DPS.  Intercept fisheries in West Greenland and St. Pierre et Miquelon, by-catch in recreational fisheries, and poaching result in direct mortality or cause stress, thus reducing reproductive success and survival.  Due to the continued involvement of the International Council for the Exploration of the Sea (ICES) and NASCO, the scientific advice from ICES to NASCO has been that there is no harvestable surplus of the mixed stock off West Greenland available for commercial harvest.  As a result, since 2002 within the West Greenland Commission of NASCO there has been agreement for no commercial export and to restrict the fishery at Greenland to an internal-use-only fishery.

Recreational angling of many freshwater species occurs throughout the range of the GOM DPS, and the potential exists for the incidental capture and misidentification of both juvenile and adult Atlantic salmon (Fay et al. 2006).  Juvenile Atlantic salmon may be easily misidentified as brook trout, brown trout, or landlocked Atlantic salmon, while adult salmon may be confused with adult landlocked Atlantic salmon or brown trout (Fay et al. 2006).  Direct or indirect mortality may result even in fish that are released as a result of injury or stress.

Mixed-stock fisheries

Commercial fisheries for Atlantic salmon within the United States have been closed since 1947; however, small but significant fisheries continue within the species’ migratory corridor off the coast of Canada and Greenland.  To effectively engage in issues requiring international collaboration such as these distant water fisheries, the United States maintains a presence at the North Atlantic Conservation Organization (NASCO) and International Conference for the Exploration of the Seas (ICES).  The United States is a signatory to the “Convention for the Conservation of Salmon in the North Atlantic Ocean” which entered into force in October 1983, creating NASCO to ensure that the burden of Atlantic salmon conservation was shared by both States of Origin and Distant Water Countries. NASCO promotes the conservation, restoration, enhancement, and rational management of salmon stocks in the North Atlantic Ocean through international cooperation.  NASCO has six members, which include Norway, the United States, European Union (EU), Canada, the Russia Federation, and Denmark (in respect of the Faroe Islands and Greenland). The United States is represented at NASCO by scientists and managers from NOAA Fisheries as well as staff from the Department of State, other Federal and non-federal agencies, and private sector advisors.   NMFS’ role is to work to reduce impacts to U.S. stocks from distant water fisheries, and seek to hold ourselves and other countries accountable for the protection and conservation of Atlantic salmon.  NMFS scientists compile and analyze data on the status of the GOM DPS and take this information to the International Council for the ICES Working Group on North Atlantic Salmon. This group takes and analyzes data from throughout the North Atlantic to provide scientific advice to NASCO. NMFS scientists coordinate and participate in the international sampling effort for the Greenland fishery.

Regulatory measures for the salmon fishery off the coast of Greenland are adopted through the West Greenland Commission of NASCO and are due to be renewed in 2015.  Since 2002, ICES has advised that there should be no fishery for Atlantic salmon off West Greenland given the precarious state of many of the contributing stocks.  Current NASCO regulatory measures provide for a fishery to occur for internal use, noting that level to be about 20 mt in the past.  When the current regulatory measures were established, Greenland allowed unlicensed fishermen to harvest salmon for personal use and licensed fishermen to sell salmon in open air markets and for use in some local institutions, such as hospitals and hotels.  In 2012, Greenland authorized up to 35 mt of landings to factories for processing and freezing, thus providing a means for expanding markets beyond the consumption-delimited fresh fish markets noted above, and providing an incentive for increasing fishing effort.  Since the adoption of the Executive Order that allowed factory landings in 2012, total reported landings of salmon in Greenland have risen steadily (35 mt in 2012, 48 mt in 2013, and 58mt in 2014).  These reported landings are in addition to unreported catch.  Greenland currently estimates unreported catch at 10 mt, but has no real basis for that amount.  For 2014, Greenland lowered the factory landings quota from 35 mt to 30 mt (noting international pressure to do so), but preliminary reports suggest this quota was exceeded by roughly 5 mt.  At present, the other components of the fishery (personal consumption, private sales, etc.) remain unlimited.

Populations of U.S. origin salmon are also harvested by St. Pierre and Miquelon (an off shore territory of France located off the coast of Newfoundland).  While smaller in scale than the West Greenland fishery, it operates outside any international management regime as France (in respect of St. Pierre and Miquelon) has refused to join NASCO.  Moreover, the domestic management regime in place does not effectively limit what can be caught.  The United States and Canada have worked together to ask France (in respect of St. Pierre and Miquelon) to join NASCO or at least cooperate more closely.  A key area of cooperation concerns implementation of an effective sampling program so the population level effects of the fishery on U.S. and Canadian origin salmon can be more clearly understood.  Progress even in this area has been slow, but there are some encouraging signs that the sampling program may be improving.

U.S. origin salmon are also harvested in Labrador.  Three user groups (aboriginal fishers, residents fishing for food, and recreational fishers) harvested a total of 44.0 t of salmon in Labrador in 2011 (summarized in a paper presented by Canada to the North American Commission in paper NAC(12)8).  The Aboriginal fisheries accounted for 89% of the total harvest (by weight) followed by 5.9% for the recreational fisheries and 4.7% for the resident’s fisheries.  The harvest totals for these users groups were similar to those reported since 2004.  Recreational fishery harvest occurred in rivers and accounted for 10% of the small salmon harvested and 0% of the large salmon harvested.  The majority of the subsistence food fishery harvest occurs in estuaries with 18.4% (7.6 t) occurring in coastal areas.  Recent genetic information presented by Bradbury et al (2014) and ICES (2014) that salmon of U.S> origin accounted for some level of harvest in these fisheries. 



Back to Top

Factor C.  Disease and predation

As summarized in the 2009 listing rule, disease and predation were lesser stressors that contributed to the historical declines of the GOM DPS.

Disease (secondary threat)

Disease outbreaks, whether occurring in the natural or hatchery environment, have the potential to cause negative population-wide effects.  Atlantic salmon are susceptible to numerous bacterial, viral, and fungal diseases.  Bacterial diseases common to New England waters include bacterial kidney disease (BKD), enteric redmouth disease (ERD), cold water disease (CWD), and vibriosis (Mills 1971; Gaston 1988; Olafsen and Roberts 1993; Egusa, 1992).  Fungal diseases such as furunculosis can affect all life stages of salmon in both fresh and salt water, and the causative agent (Saprolignia spp.) is ubiquitous to most water bodies.  The risk of an epizootic event occurring during fish culture operations is greater because of the increased numbers of host animals reared at much higher densities than would be found in the wild.  A number of viral diseases that could affect wild populations have occurred during the culture of Atlantic salmon, such as infectious pancreatic necrosis, salmon swimbladder sarcoma virus, infectious salmon anemia (ISA), and salmon papilloma (Olafsen and Roberts 1993). ISA is of particular concern for the GOM DPS because of the nature of the pathogen and the high mortality rates associated with the disease.  Most notably, a 2001 outbreak of ISA in Cobscook Bay led to an emergency depopulation of all commercially cultured salmon in the bay.

Parasites can also affect salmon.  There are more than 30 parasites of Atlantic salmon, of which the sea louse is one of the more common (Fay et al. 2006).  Sea lice are common and infestations can occur on wild fish in some areas where Atlantic salmon farming is concentrated (Fay et al. 2006).  A severe infestation can result in scale loss and flesh exposure.  Furthermore, sea lice may be vectors for diseases, particularly ISA (Nylund et al. 1994) as lice move from fish to fish (Fay et al. 2006).

Federally managed conservation hatcheries adhere to rigorous disease prevention protocols and management regulations; these must be continued.  These protocols and regulations are designed to prevent the introduction of pathogens into the natural and hatchery environments; prevent and control, as necessary, disease outbreaks in hatchery populations; and prevent the inadvertent spread of pathogens between facilities and river systems.

Predation (secondary threat)

As summarized in the 2009 listing rule, the impact of predation on the GOM DPS is important because of the imbalance between the very low numbers of adults returning to spawn and the increase in population levels of double-crested cormorants, striped bass, seals, and nonnative predators, such as smallmouth bass.  Increasing levels of predators combined with decreasing abundance of alternative prey sources has likely increased predation mortality on juvenile Atlantic salmon, especially at the smolt life stage.

Back to Top

Factor D.  Inadequacy of existing regulatory mechanisms

Regulatory Mechanisms Related to Dams (significant threat)

Atlantic salmon require access to suitable habitat to complete their life history.  As described under Factor A, dams within the range of the GOM DPS impede access to much of the suitable habitat that was historically available.  The turbines of power generation also cause significant mortality of smolts as they migrate past dams on their journeys to the ocean.  Turbine mortality can result from entrainment in racks designed to exclude fish from turbine bays, or when fish pass through the turbines themselves because the racks have not been properly designed or installed. 

As explained in the 2009 rule, hydroelectric dams in the GOM DPS are licensed by FERC under the Federal Power Act (FPA).  There are 19 FERC-licensed dams in the Androscoggin watershed (16 are impassable), 18 in the Kennebec watershed (15 are impassable), and 23 in the Penobscot (12 are impassable).  It should be noted that the numbers of impassible dams may include dams with fish ladders that may have been abandoned over the years, as well as some dams that may be considered impassible because there are no salmon in the ecosystem, or fish may be trapped at a lower watershed dam and trucked up past the impassible dams.  In these cases, the impassible dams are not the causes of mortality.

Current licenses for many dams, though not all, contain a reservation provision under FPA section 18 (16 U.S.C. 797) that could allow fishways to be prescribed by the Services (16 U.S.C. 811).  However, exercise of that authority requires administrative proceedings before the FERC and the Services.  The FERC maintains that, for the remainder of the projects whose licenses do not contain reserve authority, reopening these licenses may be dependent upon the success of a petition to the FERC to exercise its own reserve authority.  The Services maintain that the listing of a species, designation of critical habitat, or the availability of any new information should trigger a reinitiation.  Regardless, if take is occurring at these facilities, such incidental take needs to be authorized under section 7 or 10 of the ESA.  The habitat degradation and ecological impacts caused by these dams cannot be addressed by the Services’ prescriptive authority under section 18 of the FPA, but may be under FPA section 10(j) (16 U.S.C. 803) recommendations.

The majority of dams within the GOM DPS do not generate electricity, and therefore do not require either a FERC exemption or a Maine Department of Environmental Protection (MDEP) water quality certification.  These dams are typically small and historically were used for a variety of purposes, including flood control, log drives, mill working, storage, recreation, and processing water.  Most of these facilities do not have fish passage, and many of them are not in use.  Before salmon were listed, lack of fish passage and other impacts to salmon could be addressed only through State law, as noted previously.  Further, although the USFWS worked in Maine on salmon passage for decades before the DPS was listed, to a large extent the fishways we approved were either not effective or were not maintained properly.  Overall, the inadequacy of existing regulatory mechanisms relating to dams is a significant threat to the GOM DPS 

Back to Top

Factor E.  Other natural or man-made factors affecting its continued existence

Low Marine Survival (significant threat)

Low Marine survival continues to be one of the top three factors, if not the most important factor, for the continued low population number for the GOM DPS and for Atlantic salmon throughout the North Atlantic, despite significant reductions in commercial intercept fisheries.  Overall, marine survival is poor throughout the Atlantic Ocean and is heavily influenced by both nearshore and open ocean survival rates (ICES 2008). 

Marine survival is indexed by smolt return rates; a smolt return rate is the ratio of the number of adult returns produced by a smolt cohort to the number of outmigrating smolts (number of naturally reared smolts and/or the number of stocked hatchery smolts).  It should be noted that, by using this method, marine survival incorporates a significant amount of mortality that may originate in the freshwater or estuarine system from dam-associated direct, indirect, or delayed mortality (see Factor A).

Using the above approximation, marine survival has declined over the last 25 years but may be cyclical (Friedland et al. 2003) as has been documented in Pacific salmon (Francis and Hare 1994).  From 1970 to the early 1980s, the yearly return rate of hatchery stocked smolts to the Penobscot River was approximately 0.7 percent, whereas from the 1990s to the present, this value was approximately 0.15 percent (USASAC 2013).  Return rates for naturally reared smolts have generally been less than 1.5 percent. The 2001 to 2005 2SW return rates for naturally reared Narraguagus River smolts ranged from 0.2 to 1.2 percent.  For the same time period, 2SW return rates for hatchery-reared Penobscot River smolts ranged from 0.03 to 0.07 percent, and for 1SW  hatchery-reared Penobscot River smolts, from 0.06 to 0.17 percent.  It is well established in the literature that naturally reared fish survive at higher rates than their hatchery-reared counterparts (Locke 1998).

Figure 12 illustrates that significant increases in freshwater and marine survival are needed to increase populations.  The dot approximates the current freshwater survival (3.5 percent) and marine survival (0.1 percent) regimes.  The two red lines bracket the range of reported freshwater survival from a healthy population with suitable habitat conditions.  The curved line represents all possible combinations of marine and freshwater survival that will result in a population that will remain at a constant size.  If population survival falls precisely on the curved line, it would be replacing itself; that is, each spawning pair would theoretically produce two adult offspring, one male and one female.  Combinations of freshwater and marine survival that are above the curved replacement line result in population growth.  From this figure, it is clear that, while likely harder to achieve, incremental increases in marine survival have a much greater potential to result in population growth than comparable increases in freshwater survival.

Without significant increases in marine survival, recovery of the GOM DPS of Atlantic salmon is unlikely.  Unfortunately, our ability to take direct actions to increase marine survival may be limited.  Much more information, analysis, and research are needed to achieve a clearer picture of marine survival and whether actions we can implement can increase it.  As long as marine survival remains low, efforts to reduce freshwater threats are even more important.

Figure 4.  Graphical representation of the combinations of freshwater and marine survival needed for replacement.

Back to Top

Depleted Diadromous Communities (secondary threat)

Dam building played an important role in reducing Atlantic salmon and other co-evolved diadromous species.  The Atlantic salmon is one of 12 native diadromous species in Maine:  Atlantic shad, alewife, blueback herring, sea-run brook trout, rainbow smelt, shortnose sturgeon, Atlantic sturgeon, striped bass, sea lamprey, and Atlantic tomcod are anadromous.  The American eel is the only catadromous species.

Conservative estimates of the historical numbers of the diadromous fishes that co-evolved with Atlantic salmon are in the millions: 1,000,000 alewives in the Penobscot in 1867; 1,200,000 alewives per year in 1 tributary of the Kennebec in the early 1800s; 2,472,000 alewives in the Damariscotta in 1896; 400,000 blueback herring in the Kennebec in 1880; 2,000,000 American shad in the Penobscot prior to the 1830s; and 3,500,000 rainbow smelt in the Penobscot in 1887 (Saunders et al. 2006).

Damming rivers, thus preventing migration to former spawning grounds, was a major factor in the decline of Atlantic salmon, sturgeon, river herring (blueback herring and alewife), and shad (Moring 2005, Limburg and Waldman 2009).  Many co-evolved diadromous species have experienced dramatic declines throughout their ranges, and current abundance indices are fractions of historical levels.  

Rangewide population declines are dramatic:  American shad, 97 percent; alewife, 99 percent; blueback herring, 99 percent; rainbow smelt, 99 percent; Atlantic sturgeon, 91 percent; Atlantic salmon, 96 percent; and American eel, 72 percent (Limburg and Waldman 2009).  Similarly, in Maine current population levels are orders of magnitude smaller than historical population levels (Saunders et al. 2006). 

This dramatic decline in diadromous species has negative impacts on Atlantic salmon populations.  At historical levels, the alewife, blueback herring, American shad, rainbow smelt, and sea lamprey likely provided several important benefits for Atlantic salmon, such as providing alternative prey for predators of salmon (i.e., prey buffering) (Saunders et al. 2006), serving as food for juvenile and adult salmon (Cunjak et al. 1998), nutrient cycling (Durbin et al. 1979, Nislow and Kynard 2009), and habitat conditioning (Saunders et al. 2006).  Habitat conditioning refers to sea lamprey, whose spawning movements remove sediment and algae from stream rocks, which may improve habitats for Atlantic salmon spawning (Garner 2012, Guyette 2013, and Hogg 2013).

Because diadromous fish populations have been significantly reduced, ecological benefits from marine derived nutrients (MDN), habitat conditioning, prey buffering, and alternative sources of food for Atlantic salmon are significantly lower today compared to historical conditions.  These impacts may be contributing to decreased survival through (1) reduction of prey for reconditioning kelts, (2) increased predation risks for smolts in lower-river and estuarine areas, and (3) increased predation risks to adults in estuarine and lower river areas.  Although these impacts do not occur in the open ocean, the demographic impact to the species occurs after smolt emigration, and is thus a component of the marine survival regime.

Removing dams and providing access to spawning habitat will significantly help restoration of all diadromous species in the GOM DPS, which should, in turn, aid in the recovery of Atlantic salmon.  The ecological interplay and interdependence of Atlantic salmon and other diadromous species is still being investigated, but at present the lack of a robust diadromous community is a lesser stressor on the GOM DPS.

Back to Top

Artificial Propagation (secondary threat)

The conservation hatchery programs at Craig Brook and Green Lake National Fish Hatcheries (NFH) are vital to preserving individual and composite genetic stocks until freshwater and marine conditions improve for wild adults to reach stable recovery numbers. Currently, progeny produced from wild and captive broodstock are released into their rivers of origin as eggs, fry, parr, and smolts.  In addition, surplus adult broodstock are returned to their river of origin. During 2012, approximately 3.4 million juvenile Atlantic salmon were stocked into GOM DPS rivers.  Approximately 1.9 million fry were stocked, along with 17,000 0-parr, 300,000 1-parr, and 670,000 1-smolts (USASAC 2013).   

Preservation of the genetic integrity of populations, and the genetic diversity within and among populations, is critical for the long-term fitness and viability of populations (Reed and Frankham 2003; Reed et al. 2003).  An inherent risk associated with the broodstock and stocking program in the GOM DPS is the risk of domestication and loss of genetic variability, in addition to the potential for catastrophic loss due to the limited number of hatcheries maintaining GOM DPS Atlantic salmon (SEI 2007).  To minimize the genetic risks associated with maintenance of captive broodstocks, a broodstock management plan (Bartron et al. 2006) was implemented, with the goal of maintaining genetic diversity throughout the hatchery management process, including estimating genetic diversity for each captive broodstock.  Numbers of fish alone, however, do not describe a recovered Atlantic salmon species.  To achieve recovery for the Atlantic salmon population, it is necessary to demonstrate that the number of wild-origin fish meet the recovery criteria.  While there may still be some hatchery program in operation, the wild component of the population must be self-sustaining and independent of a hatchery program, if one is still operating for other purposes.  These essential characteristics are descriptive of a population that has stabilized at a robust level which provides confidence in the ability of that population to contend with natural variability.

Back to Top

Aquaculture (secondary threat) 

As summarized in the 2009 listing rule, concerns about aquaculture continue, but recent advances in containment and marking of aquaculture fish offer more control over these threats and reduce the risk of negative impacts of aquaculture fish on the GOM DPS. Aquaculture poses a disease risk by exposing native salmon in the GOM DPS to serious salmon pathogens.  Permits issued by the Army Corps of Engineers (ACOE) and MDEP require genetic screening to ensure that only North American-strain salmon are used in commercial aquaculture; marking to facilitate tracing fish back to the source and cause of the escape; containment management plans and audits; and rigorous disease screening.  Aquaculture is a lesser stressor on the GOM DPS; however these measures do not eliminate the risk aquaculture fish pose to wild Atlantic salmon but serve to reduce the potential for negative impacts.  It is important to note that at this time equally protective requirements regarding salmon aquaculture do not exist on the Canadian side of the border.  Fish held in Canadian cages, or those that may escape from Canadian cages, can still pose disease, genetic, and ecological risks to U.S. Atlantic salmon.

Back to Top

Competition (secondary threat)

Prior to 1800, the resident riverine fish communities in Maine were made up of native species, consisting of brook trout, cusk (burbot), white sucker, sculpin, and a number of minnow species.  Today, Atlantic salmon coexist with a diverse array of nonnative resident fishes, including brown trout, largemouth bass, smallmouth bass, and northern pike.  The range expansion of these nonnative species is of particular concern since they often require similar resources as salmon and therefore are expected to be competitors for food and space.  Competition is a lesser stressor to the GOM DPS because it can exclude salmon from preferred habitats, reduce food availability, and increase predation.

Climate Change (emerging threat)

Fay et al. (2006) and NRC (2003) summarize the potential impacts of climate change on Atlantic salmon. At the time of listing in 2009, although there was reasonable certainty that climate change was affecting Atlantic salmon in the GOM DPS, there was uncertainty surrounding specifically how and to what extent.  Since listing, new and emerging science has helped us gain a better understanding of these effects and just what the ramifications are for salmon.  Recent information indicates that climate change is having significant impacts on the ecosystems that Atlantic salmon depend on and subsequently having significant impacts on the overall survival and recovery of Atlantic salmon (Mills et al. 2014). Following is a synopsis of the effects of climate change, and the new and emerging science that has elevated its concern for Atlantic salmon. 

Since the 1970s there has been a historically significant change in climate (Greene et al. 2008). Climate warming has resulted in increased precipitation, river discharge, and glacial and sea-ice melting (Greene et al., 2008). The past 3 decades have witnessed major changes in ocean circulation patterns in the Arctic, and these were accompanied by climate associated changes as well (Greene et al., 2008). Shifts in atmospheric conditions have altered Arctic ocean circulation patterns and the export of freshwater to the North Atlantic (Greene et al. 2008, IPCC 2006). With respect specifically to the North Atlantic Oscillation (NAO), changes in salinity and temperature are thought to be the result of changes in the earth’s atmosphere caused by anthropogenic forces (IPCC 2006). 

Global climate change can affect all aspects of the salmon’s life history as entire ecosystems can shift rapidly (compared to evolutionary timescales) from one state to another, altering habitat features through increases in sea surface temperatures (IPCC 2001); changes in frequency of seasonal cycles of phytoplankton, zooplankton and fish populations in the marine environment (Greene and Pershing 2007); changes in freshwater hydrologic regimes; and altering the timing and frequency of river ice flows.  All of these factors can significantly alter the ecosystem in which salmon have become adapted by effecting environmental cues that stimulate migration, spawning and feeding activities. 

Friedland et al. (2005) summarized numerous studies that suggest that climate mediates marine survival for Atlantic salmon as well as other fish species. Recent analyses of bottom water temperatures found that negative NAO years are warmer in the north and cooler in the Gulf of Maine (Petrie 2007). Positive NAO years are warmer in Gulf of Maine and colder in the north (north of 45° N) (Petrie 2007). Strength of NAO is related to annual changes in diversity of potential predators: at southern latitudes, there are more species during positive NAO years (Fisher et al., 2008). The effect is system-wide where 133 species showed at least a 20 percent difference in frequency of occurrence in years with opposing NAO states (Fisher et al. 2008).

In a recent study, Mills et al. (2014) was able to associate a major decline in Atlantic salmon abundance to a series of oceanic changes across multiple levels of a salmon’s ecosystem as a result of changing climate conditions.   Her results suggest that climate driven environmental factors and warmer ocean temperatures resulted in poor trophic conditions constraining the productivity and recovery of Atlantic salmon populations in the North Atlantic.  Though all Atlantic salmon in the North Atlantic are affected by the changes in trophic conditions, the effects on populations dominated by 2 sea-winter fish (such as the GOM DPS) appears greater than populations dominated by 1 sea-winter fish.  This suggests that there is a greater cumulative effect of poor trophic conditions on 2 sea-winter fish as a result of longer residence times at sea.  Mills’ study goes on to suggest that the impacts to Atlantic salmon are most associated with salmon’s ecosystem response to warming rather than the direct impacts of warming itself.   These effects include changes to phytoplankton and zooplankton communities that salmon’s principle prey species, capelin, feed on.   Subsequent to these changes, the size, distribution and behavior patterns of capelin has shifted making them less available for salmon to prey on, subsequently reducing the overall fitness and survival of Atlantic salmon.

Within the freshwater range of Atlantic salmon, water temperature is one of the most important environmental factors affecting all forms of aquatic life in rivers and streams (Annear et al. 2004).  In addition to climate change, water temperature can be influenced by changes in riparian cover, dams, alterations in stream channel morphology (Annear et al. 2004), waste water discharge, and urban development.  Among rivers within the GOM DPS, records extending back to the early 1900’s indicate that spring runoff has become earlier, fall ice-on is later, and there are fewer days of total winter ice on Maine rivers (Dudley and Hodgkins 2002).  In support of these observations, a combination of land-surface and sea surface air temperature data shows an overall increasing trend in annual air temperatures for New England between the period of 1901 to 2000, with the greatest seasonal warming rates occurring in the winter months December, January and February as indicated by a period of record extending from 1976 to 2000 (IPCC 2001).  Several studies indicate that small thermal changes may substantially alter reproductive performance, species distribution limits, and community structure of fish populations (Van Der Kraak and Pankhurst 1997, McCormick et al. 1997, Keleher and Rahel 1996, McCarthy and Houlihan 1997, Welch et al. 1998, Schindler 2001).  Changes in fish community structure can alter predator/prey assemblages by decreasing qualitative habitat features that benefit salmon while concurrently increasing habitat features that benefit predators and competitors. 

Temperature is especially important for Atlantic salmon given that they are poikilothermic (i.e. their body temperatures and metabolic processes are determined by temperature).  Temperature can be a stimulant for salmon migration, spawning, and feeding (Elson 1969).  Thermal changes of just a few degrees celsius can critically impact biological functions of salmon including: metabolism (McCarthy and Houlihan 1997; Somero and Hofmann 1997; Reid et al. 1998), reproductive performance (Van Der Kraak and Pankhurst 1997), response to contaminants (Reid et al. 1997) and smolt development (McCormick et al. 1998).  Unnatural changes in water temperatures may also affect growth, survival and migration timing of Atlantic salmon in freshwater, the survival and timing of migrating smolts in the estuarine environment, and the survival of juveniles soon after entering the marine environment (NRC 2003).  Juanes et al. (2004) examined migration timing data from the Connecticut River drainage and from drainages in Maine and Canada and found a shift towards earlier peak migration dates across systems, correlating with long term changes in temperature and flow that may represent a response to global climate change.  For migrating smolts, the interrelatedness of water temperature and photoperiod may be extremely important to consider.  One of the concerns with climate change is the rate at which water temperatures increase could conceivably regulate the window of opportunity in which smolts can successfully transition from freshwater to saltwater.  McCormick et al. (1998) suggested that smolts experiencing delays in migration, such as those that occur at dams, may have lower survival rates if they are unable to reach saltwater within the migration window.  One possible explanation for this reduced survivorship is that a shortened migration window due to increased temperatures could conceivably result in increased predation pressure as more smolts are forced to migrate over a shorter period of time. 

Back to Top

Return to Recovery Plan Outline

Document Actions