Atlantic salmon (Salmo salar) COSEWIC assessment and status report: chapter 9


Atlantic Salmon have complex and plastic life histories that begin in freshwater and may involve extensive migrations through freshwater and marine environments before returning to fresh water to spawn.

Freshwater Habitat Requirementsvii

Atlantic Salmon rivers are generally clear, cool and well oxygenated, with low to moderate gradient, and possessing bottom substrates of gravel, cobble and boulder (COSEWIC 2006b).

Habitat is considered a limiting resource to freshwater production and is used to set conservation requirements for Canadian rivers (O’Connell et al. 1997a). Loss of freshwater habitat since European colonization has resulted in dramatic declines in the range and abundance of Atlantic Salmon (Leggett 1975). A relatively small but locally significant amount of habitat has been created by enhancing passage through the removal of natural barriers. This has increased salmon population size in several rivers (e.g. Mullins et al. 2003).

Freshwater habitat use by Atlantic Salmon is diverse, widely documented and the subject of substantial reviews (Bjornn and Reiser 1991, Gibson 1993, Bardonnet and Bagliniere 2000, Armstrong et al. 2003a, Rosenfeld 2003, Amiro 2006). Spawning beds are often gravel areas with moderate current and depth (Fleming 1996), but habitats used by juvenile and adult salmon range across freshwater fluvial, lacustrine and estuarine environments. Individual fish may often use several habitat types during their freshwater residency (Erkinaro and Gibson 1997, Bremset 2000) for demographic (Saunders and Gee 1964), and ecological reasons (Morantz et al. 1987, Bult et al. 1999).

Juvenile salmon typically maintain relatively small feeding territories in streams, which can be relocated when individuals undergo larger-scale movements to seek improved foraging conditions, refuge (thermal or seasonal) and/or precocious spawning (McCormick et al. 1998). In some areas (e.g. Newfoundland), juveniles also occupy lacustrine habitats where growth benefits are accrued (Hutchings 1986). In winter, parr may occupy interstitial spaces in the substrate (Cunjak 1988) and/or move to lacustrine habitats (Robertson et al. 2003). Ultimately, home ranges in freshwater are abandoned when smolt begin to migrate to the marine environment (the Lake Ontario populations, which likely migrated to lake environments, were an exception to this generalization). The propensity for migration underscores the importance of habitat connectivity, not only to allow adults to reach spawning grounds, but also for seasonal movements of juveniles and ontogenetic shifts in habitat.

In Lake Ontario, adult ‘Lake’ salmon typically remained in the lake until immediately prior to spawning, at which time they ascended their natal streams and established spawning sites. The small size of most tributaries of Lake Ontario and their low flow and volume were, in most cases, unfavourable for the extended residency of large salmon (Parsons 1973). Adults rarely remained in the streams longer than one week after spawning (Parsons 1973). Little is known about the preferred lacustrine habitat of Atlantic Salmon except that lakes with deep, cool, oligotrophic conditions, a forage base that includes rainbow smelt (Osmerus mordax), and the presence of feeder streams providing suitable spawning and nursery habitat, appear to be the most ecologically suitable (MacCrimmon and Gots 1979, Cuerrier 1983). Historically, Lake Ontario salmon may have depended on cisco and later alewife before smelt entered the lake in the 1930s. Lake Ontario most likely served the same function for adult and juvenile lake salmon as the ocean did for anadromous populations.

Chemical conditions also play a role in defining salmon habitat. Atlantic Salmon populations can experience reduced production or even extirpation in conditions of low pH (DFO 2000). Tolerance is life-stage dependent with fry and smolt being the most sensitive. Generally rivers that have pH’s between 4.7 and 5.0 are considered moderately impacted and those below 4.7 are considered acidified (DFO 2000), and are unlikey to be able to support salmon populations.

Temperature has been described as the most pervasive abiotic attribute controlling the production of teleost fishes in streams (Heggenes et al. 1993). Relative to other salmonids, Atlantic Salmon parr are relatively tolerant of high water temperatures (Elliot 1991). Temperatures above 22°C are unsuitable for feeding (Elliot 1991) and the maximum incipient lethal temperature (the temperature at which all salmon would exit a habitat if the opportunity were available) was estimated to be 27.8°C (Garside 1973). There is a gradual increase in smolt age associated with increasing latitude which is considered to depend upon growth opportunities in spring and summer (Metcalfe and Thorpe 1990). Therefore, it is entirely possible that an optimum temperature regime exists, affecting Atlantic Salmon abundance via smolt productivity.

Available habitat is a direct function of discharge (Bovee 1978) and exposure of juvenile populations to extended low flow periods may limit production in streams. Low flows have also been widely observed to delay entry of returning spawners to freshwater environments (Stasko 1975, Brawn 1982). Variation in flow, however, is normal in the temperate streams that salmonids occupy. Atlantic Salmon have been noted for their capacity to cope with this variation in flow and associated physical constraints relative to other sympatric salmonids. Juvenile salmon were noted to move from pool to riffle habitats at higher discharges (Bult et al. 1999), which is complementary to the noted preference of pools at low discharge (Morantz et al. 1987). This adaptability enables juvenile salmon to occupy extensive sections of streams that experience flow and temperature variation.

The migratory behaviour exhibited by Atlantic Salmon makes them particularly vulnerable to the negative effects of obstructions. Both natural and man-made barriers to fish passage severely reduce the production of salmon by restricting mature salmon from reaching spawning habitat and preventing juveniles from reaching feeding and refuge habitats. In general, most obstructions in excess of 3.4 m in height will block the upstream passage of adult salmon (Powers and Orsborn 1985). Ideally, a passable falls will have a vertical drop into a plunge pool with a depth 1.25 times the height. Depending on the shape of the falls and plunge pool, the maximum height can be considerably less. Furthermore, since jumping and swimming capacity is a function of body length (Reiser and Peacock 1985), the ability of juveniles to surmount barriers is greatly reduced relative to adults.

Marine Habitat Requirementsviii

Salmon move, as juvenile smolts or post-spawning ‘kelts’, from fresh water to brackish estuaries and then to the open ocean (Figure 11). O’Connell et al. (2006) report that it is in the ocean where “growth… is rapid relative to that in fresh water… mass increases about 75-fold between the smolt stage and 1SW salmon stage, and over 200-fold from smolts to 2SW salmon”. Overall natural mortality in the sea is high and variable and there are many factors that can affect the survival of Atlantic Salmon, some habitat-related (Reddin 2006). However, Reddin (2006) also reports “population-specific information is lacking concerning the cause of these mortalities and this is partly because detailed information on migration routes and distribution is generally unavailable for specific populations, although it is thought that their distributions generally overlap in the North Atlantic.

Survival rates associated with the transition from fresh water to ocean life for Atlantic Salmon, whether for smolts or kelts, have an important influence on year-class strength (Reddin 2006). It is generally thought that water temperature is the main controlling environmental variable for smoltification (although photoperiod is also important). The smolt transformation process is accompanied by changes in metabolic rate, with increases in energy demands underpinning the need for the fish to immediately begin feeding. Of all the variables influencing survival of ‘postsmolt’ (individuals experiencing their first several months at sea) salmon, temperature is particularly important because temperature regulates metabolic rate. If postsmolts are to survive, individuals must quickly adapt to their new physical environment and be able to escape predators and capture prey. Temperatures occupied by salmon range from below 0 to nearly 20°C, although most were 8-15°C (Reddin 2006). The length of time spent in or near the home estuary is thought to be as brief as 1-2 tidal cycles and may limit opportunities for predation. In general, postsmolt movement to oceanic areas is rapid. Tracking studies confirmed this rapid movement away from estuaries towards the open sea and showed that migration was influenced by tidal currents and wind (Hedger et al. 2008; Martin et al. 2009). One exception was in the Gulf of St. Lawrence where salmon postsmolts were caught in a nearshore zone late in the summer; presumably long after they had left their home river and estuary (Dutil and Coutu 1988). In North America, movement of postsmolts, once in the open sea, is generally northwards.

Research surveys for postsmolts in the Northwest Atlantic have yielded highest catches and catch rates between 56° and 58° N in the Labrador Sea; capture dates and behaviour suggest that some postsmolts probably overwinter there as well (Reddin 2006). Postsmolts in the Labrador Sea originate from rivers over much of the geographical range of salmon in North America, but the degree of their migration to the Labrador Sea varies by population. Postsmolts have also been caught as bycatch in herring gear in the northern Gulf of St. Lawrence in late summer. The winter destination of these salmon remains unknown. Postsmolts from rivers in the inner Bay of Fundy have been observed to remain in the Bay of Fundy until late summer. Although the overwinter location of iBoF salmon is unknown, the lack of tag recoveries from distant intercept fisheries indicates that iBoF salmon do not go as far north as other salmon stocks.

In spring, adult salmon are generally concentrated in abundance off the eastern slope of the Grand Bank and less abundantly in the southern Labrador Sea and over the Grand Bank. During summer to early fall, adult, non-maturing salmon are concentrated in the West Greenland area and less abundantly in the northern Labrador Sea and Irminger Sea. There are notable exceptions to these tendencies. As for postsmolts from the same area, few adult salmon from the iBoF are caught outside the Bay itself. Another exception is Ungava Bay, where salmon from local rivers are known to overwinter. In some cases adults from ‘spring run’ populations may be migrating up-river while other conspecifics from nearby populations are well out to sea.

Sea surface temperature (SST) and ice distribution control run timing and distribution in the Northwest Atlantic (Reddin 2006). Salmon are found at sea in water with SSTs of 1-12.5°C, with peak abundance at SSTs of 6-8°C. In the Labrador Sea, 80% of the salmon were found in SSTs between 4-10°C (Reddin 2006). Similarly, tagged Atlantic Salmon kelts were found in temperatures ranging from a low near 0°C to over 25°C, although most of the time kelts stayed in seawater of 5-15°C (Reddin et al. 2004). Lethal temperatures for adult salmon occur below 0°C (Fletcher et al. 1988). This may explain the tendency of salmon to avoid ice-covered water as reported by May (1973). The significant relationship for SSTs and salmon catch rates suggests that salmon may modify their movements at sea depending on SST.

Lethal seawater temperatures for both wild and farmed salmon smolts adapting to seawater occurred at both low and high temperatures (Sigholt and Finstad 1990, Handeland et al. 2003). At the lower end of the temperature range, mortalities of postsmolts occurred at sea temperatures of 6-7°C while at the higher end, mortalities occurred at temperatures over 14°C. This suggests that there may also be environmental windows for successful smolt transition into the sea.

Friedland (1998) reviewed ocean climate influences on salmon life history events including those related to age at maturity, survival, growth and production of salmon at sea. He concluded that ocean climate and ocean-linked terrestrial climate events affect nearly all aspects of salmon life history. For example, higher sea surface temperature has been implicated in increasing the ratio of grilse to MSW salmon (Saunders et al. 1983, Jonsson and Jonsson 2004), perhaps through growth rates (Scarnecchia 1983). Also, Scarnecchia (1984), Reddin (1987), Ritter (1989), Reddin and Friedland (1993), Friedland et al. (1993), Friedland et al. (1998, 2003a, 2003b), and Beaugrand and Reid (2003) showed significant correlations between salmon catches/production and environmental cues, including those related to plankton productivity.

Figure 11: Routes of Marine Migration of Postsmolt (left panel) and Returning Adults (right panel)

Maps showing marine migration routes of postsmolt (left panel) and returning adult Atlantic Salmon (right panel).

Figure modified from Reddin (2006).

Freshwater Habitat Trendsix

Dams, with and without fish passages, probably account for the majority of salmon habitat lost in North America. Prior to the development of hydroelectric power there were extensive small mill dams. From 1815 to 1855 more than 30 mills a year were being built in the Atlantic provinces (Dunfield 1985). In Nova Scotia alone, there were a total of 1,798 dams in 1851. In both Nova Scotia and New Brunswick, surveys documented severe habitat loss and destruction caused by dams and mill waste. Estimates made at the time indicated that 70-80% of the habitat for salmon was affected. A similar situation was occurring in ‘Upper Canada’ at this time and by 1866, salmon in many tributaries of Lake Ontario were severely depleted or extirpated (Dunfield 1985).

With the development of the Fisheries Act, shortly after confederation in Canada, some habitat conditions improved. However, a new trend of development began for hydroelectricity in the late 1920s. This technology required the construction of high-head concrete dams that flooded vast areas of rivers. Fish passage structures, when installed, proved to be difficult to operate effectively and in many cases were eventually abandoned due to the lack of fish. Many of the major rivers were developed for hydroelectric power over the next 40 years and more salmon populations were lost. Because hydro developments were often associated with existing falls, not all hydroelectric power developments directly caused the loss of salmon populations. No complete inventory of dams and habitat loss is found in the literature. However, it is notable that five of the largest rivers in Nova Scotia, all of which had salmon prior to European colonization, were subsequently developed for hydropower and no longer have indigenous salmon populations (DFO and MRNF 2008). This observation is clearly not unique to Nova Scotia. Gains in habitat, though modest compared to losses, were achieved by providing passage around natural barriers. For example in Newfoundland, enhancements from the 1940s to the 1990s opened up over 21,600 ha of fluvial habitat to salmon (Mullins et al. 2003).

Overall, prior to 1870 as much as 50% of the habitat, or the populations that used those areas, were lost. The majority of these populations and areas were in the Upper St. Lawrence and Lake Ontario (Leggett 1975). The net loss of productive capacity by 1989 was estimated at 16% since 1870, 8% due to loss in productive capacity, 7% due to impoundment, and 3% due to acidification (Watt 1989). During the same period, there was a 2% increase from fish passage development (Watt 1989).

In addition to reductions in habitat availability, freshwater habitat quality has suffered in some areas due to acidification. North American emissions of SO2 increased during the industrial revolution and peaked in the early 1970s. Approximately 60% of wet sulfate deposition is from human activities in North America. Reductions in emissions have since been achieved and are reflected in both wet sulfate depositions and hydrogen ion concentrations at monitored sites. Anthropogenic sulfate deposition has decreased about one-third since the mid-1980s (DFO 2000). This has caused a large decrease in the deposition of acidifying substances. Unfortunately, the reduction in atmospheric hydrogen (H+) deposition has not resulted in a substantial decrease in lake acidity at negatively affected sites in Nova Scotia. Furthermore, reduction in acid deposition has not been reflected in the acid neutralization capacity (ANC). As a result, 22% of the 65 salmon rivers on the Southern Upland are ‘acidified’ and are known to have lost their salmon populations (DFO 2000).

There have been recent efforts to restore habitat in and around traditional salmon spawning streams, particularly in riparian areas, in the Lake Ontario drainage. It is important to note that continued increase in urbanization (and associated increase in impervious cover) of the Greater Toronto Area is likely to have direct and indirect impacts on the chemical and biological characteristics of streams in the region (Stanfield and Kilgour 2006, Stanfield et al. 2006). Within the lake itself, there have also been many changes that may negatively affect Atlantic Salmon survival including the introduction of Pacific salmon and other non-native salmonid species (Christie 1973, Scott et al. 2003), and the invasion of Lake Ontario by species such as Sea Lamprey (Petromyzon marinus) (Christie 1972) and dreissenid mussels.

Quebec and Atlantic populations are also facing varying degrees of changing land-use patterns (e.g. urbanization, forestry, agriculture) and threats from invasive species. These are qualitatively outlined in the Threats and Limiting Factors section.

Marine Habitat Trendsx

Climate change is a critical issue for Atlantic Salmon, as it can alter productivity and cause ecological regime shifts (Hare and Francis 1995, Steele 2004, Beamish et al. 1997). In the northwest Atlantic, there is evidence that a basin-scale shift (as a consequence of changes in the North Atlantic Oscillation Index) has negatively affected the productivity of Atlantic Salmon (Reddin et al. 2000, Chaput et al. 2005), and may be linked to downturns in salmon abundance (Dickson and Turrell 2000) and recruitment (Beaugrand and Reid 2003, Jonsson and Jonsson 2004, Chaput et al. 2005) in the North Atlantic. Recent research has also suggested that there may be substantial impacts on early growth in the marine environment as a consequence of climate change (Friedland et al. 2005, 2006, 2009).

Recent downturns in Atlantic Salmon abundance in the late 1980s and 1990s are unprecedented in magnitude and have drawn attention to the lack of knowledge of salmon ecology during the marine phase (Reddin 2006). Because declines in salmon abundance have been widespread, and because apart from DUs 14-16, there have been few indications of reduced smolt production in fresh water, it has been concluded that the main cause lies within the ocean phase (Reddin and Friedland 1993, Friedland et al. 1993). For many rivers where marine survival has been measured, the lowest recorded values have occurred in recent years. These low survivals have coincided with greatly reduced marine exploitation (fishing) achieved through massive reductions in effort or in some cases complete bans (ICES 2005), leaving the conclusion that something other than fishing is the main cause. Beaugrand and Reid (2003) have detected large-scale changes in the biogeography of calanoid copepod crustaceans in the northeast Atlantic in relation to sea surface temperature. It seems that copepod assemblages associated with warm water have shifted about 10° latitude northwards. Declines in a number of biological variables, including salmon abundance, have shown to be correlated with these changes (DFO and MRNF 2008). This regional temperature increase therefore appears to be an important factor driving changes in the dynamics of northeast Atlantic pelagic ecosystems with possible consequences for biogeochemical processes, all fish stocks, and fisheries. Regime shifts associated with climate change are predicted to continue, particularly in the Labrador Sea; now considered to be the “centre of action of climate change in the North Atlantic for the 21st century” (Dickson et al. 2007 in Green et al. 2008).

Unlike other populations in Canada, inner Bay of Fundy (iBoF) salmon are thought to overwinter in the Bay of Fundy / Gulf of Maine. Nonetheless, poor marine survival remains the primary driver of the collapse of iBoF stocks. Significant declines in marine habitat quality and abundance in this region may be occurring due to at least three mechanisms. First, over 400 tidal barriers have been constructed in the Bay of Fundy, and while their placement predates 1970 (Wells 1999), it is possible that cumulative effects through time have negatively altered the iBoF ecosystem for salmon. Second, a large aquaculture industry has grown in the western Bay of Fundy, northern Gulf of Maine, and southwest region of the Scotian Coast in the past 30 years. Third, primary production is apparently declining in parts of the western North Atlantic (Gregg et al. 2003). This decline might cause dramatic changes in energy flow, fish physiological condition and fish community structure, as recently indicated for the eastern Scotian Shelf (Choi et al. 2004). Potential causes of the decline in primary production include climate change (Drinkwater et al. 2003) and enormous removals of fish biomass by marine fisheries that cannot be matched by net primary production (Choi et al. 2004).

Habitat Protection/Ownership

All or part of 36 salmon rivers occur within the federally protected lands of National Parks (Terra Nova National Park DU 3: 9 rivers; Gros Morne National Park DU 6: 10 rivers; Kouchibouguac National Park DU 12: 4 rivers; Cape Breton National Park DU 13: 11 rivers; Fundy National Park DU 15: 2 rivers; Kejimkujik National Park and Historic Site DU 14: 1 river). Each national park contains only a small proportion of individuals within the corresponding DU and in some cases local populations are extirpated (e.g., Mersey River of Kejimkujik National Park and Historic Site). All remaining rivers flow through lands that are privately or provincially owned.

The federal government’s constitutional responsibilities for sea coast and inland fisheries are administered via the Fisheries Act. The Act provides Fisheries and Oceans Canada (DFO) with powers, authorities, duties and functions for the conservation and protection of fish and fish habitat (as defined in the Fisheries Act) essential to sustaining commercial, recreational and Aboriginal fisheries. The Fisheries Act contains provisions that can be applied to regulate flow needs for fish, fish passage, killing of fish by means other than fishing, the pollution of fish-bearing waters, and harm to fish habitat. Environment Canada has been delegated administrative responsibilities for the provisions dealing with regulating the pollution of fish-bearing waters while the other provisions are administered by DFO.


vii Elements of this section have been copied, abstracted and/or synthesized from DFO (2000), Amiro (2006), COSEWIC (2006a, 2006b) and DFO and MRNF (2008).
viii Elements of this section have been copied, abstracted and/or synthesized from Reddin (2006), COSEWIC (2006a, 2006b) and DFO and MRNF (2008).
ix Elements of this section have been copied, abstracted and/or synthesized from DFO (2000), Amiro (2006), COSEWIC (2006a, 2006b) and DFO and MRNF (2008).
x Elements of this section have been copied, abstracted and/or synthesized from COSEWIC (2006b), DFO and MRNF (2008) and DFO and MRNF (2009).

Page details

Date modified: