Temperature change beyond the range of thermal regimes observed in natural waters in British Columbia is expected to have the most serious impact on salmonid fisheries in freshwater environments in comparison to other water uses. To this end, an updated review of the available scientific literature on temperature criteria for individual life history stages of native species is provided to establish acceptable limits. Minimum and maximum endpoint values are tabulated for embryo/alevin, juvenile and adult stages with temperature criteria provided for incubation, rearing, migration and spawning activities. The summary borrows heavily from an extensive literature review and synthesis of freshwater temperature alterations on salmonid life history stages after McCullough (1999) and incorporates effects of rate of change (i.e., increase or decrease) where data are currently available for individual species.
4.1 Incubation
The rate of embryo and alevin development is strongly influenced by water temperature and the time to emergence is governed by changes in ambient conditions. In general terms, the higher the temperature within acceptable limits, the faster the rate of embryogenesis (Bjornn and Reiser 1991). Salmonids have evolved to complete certain life history functions within the range of ambient temperatures provided in nature. In this regard, the eggs of fall spawners experience a slower rate of development through the winter while the eggs of spring spawners experience a faster rate of development during late spring-early summer. A critical requirement of embryo development, however, involves a stable temperature regime immediately following spawning since eggs are highly temperature-sensitive soon after fertilization (Peterson et al 1977). For example, cutthroat trout eggs initially fertilized at a temperature of 7 degrees C and cooled at a rate of 1 degrees C/h had a significantly lower survival in less than or equal to 11 d than control eggs reared at 7 degrees C (Hubert and Gern 1995). More importantly, the same experiment repeated 13 d after fertilization (i.e., a stage in development following the period of neural tube closure) revealed no difference in survival between control and treatment groups. Notwithstanding, the rate of temperature decrease does not seem to be as important as the endpoint temperature value. In a related study on temperature reduction during early cutthroat, rainbow and brown trout egg incubation, a thermal shock applied at a rate of 0.25, 1, 5 and 20 degrees C/h, that lowered the initial incubation temperature from 7 to 2 degrees C, showed little difference in embryo survival among the experimental rates of adjustment tested (Stonecypher et al. 1994). The final endpoint of 2 degrees C caused the highest mortality observed as late as the eyed stage of development. Heat shock of +6.8 degrees C to incubating coho eggs acclimated to 1.5, 3.5, 4.0, 6.1 and 10.2 degrees C, over an 8 h period, only caused a reduction in survival to 50% hatching by egg batches cultured at 10.2 degrees C (Tang et al. 1987). The thermal shock applied to egg batches incubated at 10.2 degrees C was the only test case where eggs were exposed to temperatures well above their threshold (i.e., 17 degrees C versus a threshold of 13 degrees C). Despite returning individual egg batches to their original incubation temperature after the 8 h exposure, the highest endpoint was implicated as the causitive factor in reduced survival irrespective of developmental stage. Optimum, maximum and minimum incubation temperatures for native salmonids (and eastern brook trout) are provided in Table 1.
4.2 Juvenile rearing
Juvenile salmonid rearing environments are highly variable throughout their early life history stages and individual species have adapted a variety of life history strategies to facilitate survival that include lotic, lentic and estuarine habitats. Individual species occupy aquatic environments with thermal regimes that vary daily, seasonally, annually and spatially, and each species has demonstrated well-defined temperature preferences and tolerances (Bjornn and Reiser 1991). The majority of juvenile salmonids display an optimum growth zone between 10 and 15 degrees C while positive growth is maintained from about 4-19 degrees C (Armour 1991). Optimal growth temperatures of 19 degrees C with continued feeding up to 23 degrees C have been demonstrated for some species in response to an unlimited food supply but food conversion efficiencies were maximized at ~19.6 degrees C (e.g., chinook salmon, Brett et al. 1982). When food is limited however, the optimum growth zone is reduced to lower temperatures to compensate for elevated respiration/growth ratios (Elliott 1981 cited in McCullough 1999). Temperature extremes as high as 22-24 degrees C and as low as 0 degrees C are considered life-threatening for salmon species (Walthers and Nener 1997) or have been shown to alter growth in other salmonid species due to differences in metabolic efficiencies at high or low temperatures (McCullough 1999). A rise in the metabolic rate of cold-blooded aquatic organisms occurs when optimum temperatures are exceeded, which in turn, increases their energy requirements. A higher food requirement for juvenile salmonids may increase the risk of predation due to increased foraging.
Behavioural changes have been most notable in response to annual temperature variation: salmonids have demonstrated an avoidance reaction to high temperatures by seeking a variety of thermal refugia (deep pools (Matthews and Berg 1997), areas of ground water intrusion (Bisson et al. 1988; Nielsen et al. 1994), cooler tributaries (Mabbot 1982), or cooler headwater reaches (Rahel and Hubert 1991). It is important to note, however, that salmonids cannot always avoid elevated temperature conditions in the wild which places individuals at higher risk to multiple stressors that can reduce the performance capacity of the fish; metabolic energy is diverted from investment activities such as growth and reproduction and directed toward activities that restore homeostasis (i.e. respiration, locomotion, hydromineral regulation and tissue repair; Wendelaar Bonga 1997). Alternatively, fish that move from coldwater refugia to warm water feeding areas may experience a temperature shock during the transition between holding and feeding stations (and vice-versa) that could increase the risk of predation during recovery (Coutant 1973).
For juvenile salmonids in stream environments however, temperature avoidance may require certain life stages to occupy sub-optimal habitat. Under high temperature extremes, juvenile salmonids become lethargic (McCullough 1999) or may display
Table 1. Optimum water temperatures
and threshold levels for eggs/alevins of salmonids species in British Columbia.
(summaries after Bjornn and Reiser 1991 and McCullough 1999).
erratic behaviour prior to loss of equilibrium (Elliott and Elliott 1995), are unable to defend individual territories (McCullough 1999) and become more vulnerable to predation (Coutant 1973).
Due to lower metabolic activity at extremely low temperatures, salmonids also exhibit reduced activity.
Under low temperature extremes, stream-rearing juvenile salmonids seek out cover in interstitial areas of the channel bed (i.e. both large rivers and small streams) or alternative habitats associated with off-channel areas (Bjornn and Reiser 1991).
In nature, diurnal temperature fluctuations are generally greatest under hot summer, low flow conditions. Temperature shifts of 15 degrees C have been reported in small streams of low volume and limited riparian cover without high juvenile mortality, if maximum daily temperatures approaching upper incipient lethal levels are short-lived and temperatures are restored within the optimal range for growth (Bjornn and Reiser 1991). Notwithstanding, a large amplitude of diel variation can be considered a threat to fish health particularly where individuals are subjected to repeated exposure to resistance temperatures over a lengthy interval. Cyclical patterns of resistance temperature exposure can lead to chronic stress that manifests in reduced overall vigor, lower disease resistance, reduced growth, impaired maturation, poor reproductive success or altered behaviour (Wendelaar Bonga 1997). Depending on the rate of temperature change that may induce a temperature shock, individuals may be at a higher predation risk during the recovery period (Coutant 1973).
The determination of lethal temperatures has largely been laboratory-derived using two distinct methods: 1) the direct transfer of fish from an acclimation temperature to high or low endpoints to establish incipient lethal temperatures (ILT) or 2) gradual heating or cooling under specified rates of temperature change from an initial period of acclimation to yield critical thermal maxima or minima (CTM;McCullough 1999). A rapid increase or decrease in temperature can result in high mortality. The severity of temperature shocks however, are directly related to the acclimation temperature preceding the change, the magnitude of the temperature change and the period of exposure (Tang et al. 1987). Experimental evidence suggests that most fish can tolerate temperature shifts of 15-18 degrees C if exposure temperatures fall within the tolerance range of individual species (Hokanson 1977 cited in McCullough 1999). Under laboratory conditions, salmonids have shown different responses to rates of temperature change. Under a declining regime, slower rates of change have allowed individuals to achieve lower temperatures prior to loss of equilibrium, suggesting that partial acclimation occurs with a more gradual rate of cooling (Becker et al. 1977). Conversely, higher resistance temperatures are achieved with a higher rate of heating prior to loss of equilibrium or death, but exposure times are accordingly reduced (examples cited in McCullough 1999). It is important to note that heat exchange is a function of body size wherein a slower rate of heat transfer occurs in larger fish owing to their larger body volume. Regardless of mass, the body temperature of fish cools more quickly than it warms and juvenile salmonids may be more vulnerable to a 1-2 degrees C decrease near temperature minima compared to a 5-10 degrees C increase that approaches temperature maxima (examples provided in McCullough 1999). In general, adult fish are much more sensitive to temperature extremes than juveniles of the same species (Spigarelli et al. 1983 cited in McCullough 1999) and both preferred temperatures and tolerance limits are lower for adults than juveniles (McCauley and Huggins 1979)
Although variable rates of temperature change together with high endpoint values have demonstrated both sub-lethal and lethal effects, it is important to recognize the effects of a cyclic regime of diel fluctuations where some part of the fluctuation extends beyond tolerance temperature ranges over a period of weeks. To this end, the thermal exposure history may induce a cumulative response that ends in mortality where fish are subjected to unusual stress loading (DeHart 1975; Golden 1976 cited in McCullough 1999).
From a seasonal perspective, temperature fluctuations above background may have an inhibitory influence on physiological processes associated with the smoltification of anadromous salmonids. Either seaward migration or smoltification can be blocked if spring temperatures exceed a threshold of 12-14 degrees C established for most native species (Hoar 1988; Johnston and Saunders 1981).
Given that temperature thresholds for salmonids have largely been derived under a controlled environment in the laboratory, the applicability of threshold values to the wild may have its limitations where artificial conditions fail to integrate a variety of dynamic processes found in nature. The interaction of a variety of environmental effects (i.e. water quality, habitat quality, inter-specific competition or predation that manifest in physiological stress) may therefore impose constraints that lower temperature tolerance limits derived under more controlled circumstances.
In consideration of the types of land or water-based developments in BC, the effects of rapid (point-source heated effluents) or gradual (passive solar heating) temperature increases in receiving environments are more likely to be observed than rapid cooling events. The degree to which passive rates of temperature change occur in the province will likely vary according to geographical location and climate; higher rates of change expected in southern latitudes and lower rates in latitudes that are more northerly.
A summary of incipient lethal temperatures and critical thermal maxima for native salmonids are provided in Tables 2 and 3, respectively.
4.3 Adult migration
Water temperature provides a physical cue to adult spawners in spring, summer or fall to signal entry into natal streams. A variety of migration strategies are exhibited among anadromous salmonids that may involve a period of summer or winter holding in large river environments prior to immigration into natal spawning tributaries. Individual species show specific temperature preferences during seasonal migrations. Preferred upstream migration temperatures for salmon are particularly important given the limited energy reserves and allotted time requirements for both migration and spawning activities. Although individual populations have evolved to compensate for annual variation in climatic conditions, prolonged high temperatures can increase the rate at which energy stores are depleted for standard metabolism (Fry 1971 cited in McCullough 1999). Thermal blockages that delay migration and exhaust energy reserves can result in complete reproductive failure; elevated stream temperatures that are sub-lethal can indirectly influence spawning success by increasing metabolic stress (i.e., reduced oxygen levels) or reducing resistance to disease (McCullough 1999; Gilhousen 1990). Susceptibility to disease among spawners has been reported when water temperatures exceed 16 degrees C (Walthers and Nener 1997). A summary of preferred migration temperatures and tolerance limits is provided in Table 4.
Table 2. Optimum growth and upper and
lower incipient lethal temperatures of juvenile salmonids in selected states
and provinces of the Pacific Northwest (after Bjornn and Reiser 1991; McCullough
1999).
Table 3. Critical thermal maxima for juvenile salmonids reported in McCullough 1999.
Table 4. Preferred migration water temperatures and upper thermal tolerances for selected salmonids in streams (after Bjornn and Reiser 1991; Mc Cullough 1999).
4.4 Spawning
Although a period of holding of up to several months can occur between initial migration and spawning, the timing of the spawning period is inextricably linked to water temperatures conducive to incubation. Declining fall temperatures usually trigger the onset of spawning, however, sufficient time must follow to ensure that an adequate period of egg development precedes winter minima. Suitable fall temperatures oscillate around 10 degrees C for most salmonid species (Bjornn and Reiser 1991). Rising water temperatures similarly provide a physical cue for spring spawners with optimum incubation temperatures remaining less than 12 degrees C during the period of embryogenesis (refer to Table 1).
Pre-spawning water temperatures within acceptable levels are highly important with respect to egg viability prior to ovulation and maternal care following egg deposition. The viability of eggs within a maturing female is often reduced if holding temperatures exceed temperature tolerances of individual species prior to spawning (e.g., 15-16 degrees C for salmon; 13-15 degrees C for trout; McCullough 1999). Alternatively, the period of nest guarding by females, that prevents the potential re-excavation of the redd by later spawning runs, may be reduced if limited energy stores are used up prematurely in response to higher metabolic costs associated with high stream temperatures. A summary of observed spawning temperatures for the majority of salmonids present in coastal and interior watersheds is provided in Table 5.
Table 5. Observed spawning temperatures and thermal tolerances of selected salmonids in streams (after Bjornn and Reiser 1991; McCullough 1999).
