Determining the upper thermal tolerance of Athabasca rainbow trout (Oncorhynchus mykiss) across naturally varying stream temperatures in the Athabasca River watershed

Stream temperature is a key driver of physiological function in ectothermic fish, and fish have clear upper and lower limits to thermal habitat use. Stream temperature increases from climate change are a major threat to coldwater taxa like Athabasca rainbow trout (Oncorhynchus mykiss; ARBT), listed as Endangered under Canada’s Species at Risk Act. To better understand their vulnerability to climate change and prioritize recovery locations, ARBT were collected in August 2021 from three streams in the upper Athabasca River, Alberta, Canada, across a gradient of thermal regimes ranging from cold to warm. Individual thermal tolerance was measured using two physiological metrics: agitation temperature (Tag) and critical thermal maxima (CTmax) using a portable streamside laboratory. Tag is a behavioural indicator of a thermal avoidance threshold, whereas CTmax is a physiological response that can be interpreted as the upper thermal limit where long term survival is reduced. Results provide strong evidence that CTmax is a plastic metric in ABRT, however Tag is not. This suggests that ARBT likely have a moderate to limited thermal plasticity and will be forced to adapt or find new habitats under warming stream temperature regimes.

Methods
Upper thermal tolerance experiments were conducted in a mobile streamside laboratory trailer to determine CTmax and Tag. Large heating tanks were filled to a predetermined volume with stream water ensuring the experimental starting temperature reflected resident thermal conditions. Throughout the experiment, fish were individually housed in small acrylic chambers with mesh ends (n = 6). Individual chambers functioned to reduce social and physical interactions between study fish while permitting continuous water exchange between the chamber and tank. Water temperature in the tank was controlled by five submersible heaters (TH-0300S titanium heaters, Finnex, Chicago, IL, USA) to gradually increase the temperature at a standardized heating rate of 0.3 °C/min (Becker & Genoway, 1979) for both CTmax and Tag trials, with temperature monitored to ensure an accurate heating rate. Two circulation pumps and air stones were also placed in the tank to ensure uniform water temperature and saturated dissolved oxygen levels. Two experimental trials were conducted daily using identical replicate tank systems to increase fish throughput.

On the morning of each experiment, six fish were removed from the instream holding bags where they were placed the day before and moved into individual chambers within the heating tank. Fish were acclimated for 2 h at a stable temperature to reduce potential stress associated with handling and a new environment. Once the acclimation period at ambient stream temperature was complete, the dynamic heating period began. Fish movement was carefully monitored for behavioural cues that define the onset of agitation. Tag was always observed before the onset of CTmax, and was defined as 5s of consistent burst swimming or aggressive turning inside individual confinement chambers (McDonnell & Chapman, 2015). CTmax was defined by a loss of equilibrium (inability to remain upright) and/or disorganized locomotory movement (Beitinger et al., 2000). Fish were immediately removed from the warm experimental tank when CTmax was observed and placed in a cool recovery tank. Fork length (in mm) and wet body mass (in g) measurements were recorded, and a fin clip was taken from the left pelvic fin of each fish for genetic analysis (n = 120). Once fish had regained equilibrium and recovered, they were released back into the stream at the site of capture. 

Rainbow Trout of hatchery origin were historically stocked throughout the Athabasca River watershed (COSEWIC, 2014). To ensure that all fish used in the data analysis were genetically pure, fin clips from all assayed fish were genetically analyzed for introgression with stocked Rainbow Trout. All protocols closely followed those of Geraldes and Taylor (2021). Tissue samples were digested and DNA was extracted and amplified to estimate the ancestry coefficient between wild and hatchery RBT groups, expressed as the proportion of the genome of each fish that was of wild origin (i.e., QH). Although five tissue samples had insufficient genetic material for analysis, the remaining 115 samples were all genetically pure with an introgression coefficient Qi > 0.95, indicating genetic introgression was rare in experimental fish caught in McPherson, Anderson, and Wampus creeks. Based on these results, it is reasonable to infer that all experimental fish (n = 120) were genetically pure ARBT.

Long-term acclimation temperature prior to measurement of thermal tolerance calculated as the mean water temperature at each site for 7 days preceding the experiment based on observed temperature logger data. However, 2021 temperature data was missing from the upstream site in McPherson Creek, and was subsequently inferred based on the water temperature relationship between up and downstream sites of McPherson Creek measured the following summer.

Data from thermal tolerance experiments was excluded if behavioural observations were deemed inaccurate or ambiguous. Three Tag data points were excluded when fish either expressed agitated behaviour from the start of the experiment, or if fish failed to express any agitated behaviour prior to CTmax. Five CTmax observations were excluded when fish demonstrated full recovery of equilibrium and activity immediately after being placed in the recovery tank, as this indicates a behavioural misinterpretation of loss of equilibrium, and that fish were prematurely removed from the experiment.