Preferred and Optimal Swimming Speeds in Rainbow Trout (Oncorhynchus mykiss) at Three Temperatures

Abstract

Sustained exercise in aquaculture is known to improve the health and growth of finfish. Implementing exercise regimes has become an increasing focus in aquaculture practice. This study examined the relationship between the preferred swimming speed (U_pref) and the optimal swimming speed (U_opt) in rainbow trout (Oncorhynchus mykiss) under non-migratory conditions typical of aquaculture environments. Using a circular raceway, rainbow trout were allowed to swim voluntarily to determine U_pref. U_opt was measured using a forced-swimming test in a swim tunnel respirometer. Experiments were conducted at three temperatures (10 °C, 15 °C, and 20 °C). The results revealed a significant difference between U_pref (1.18 ± 0.14, 1.17 ± 0.19, and 1.24 ± 0.15 BL s⁻¹, respectively) and U_opt (1.4 ± 0.19, 1.5 ± 0.15, and 1.6 ± 0.24 BL s⁻¹, respectively) across all temperatures. Aerobic scope was greatest at 15 °C (3.8), consistent with the species’ thermal range. Notably, swimming at U_pref required 18–22% less energy than U_opt, suggesting that U_pref is more suitable for aquaculture systems. This study introduces a minimally invasive and stress-free method for determining U_pref and provides insights that can optimize flow regimes in aquaculture tanks, improving both energy efficiency and fish welfare.

1. Introduction

Swimming of fishes is fundamentally connected to biological processes such as early development, maturation, reproduction, and growth. Some species have adapted to take on long-distance migration between feeding and breeding grounds [1,2]. Many fish species have developed swimming techniques that demand minimal energetic cost and are optimized to take on long-distance migrations. Understanding these swimming dynamics and physiological traits is of great interest when we look towards domestication and industrialization of fish for aquaculture [3].

Exercising fish has been shown to improve cardiac performance, disease tolerance, and growth efficiency [4,5,6,7]. However, to implement exercise regimes in tanks or raceways in Recirculated Aquaculture Systems (RAS), a broader understanding of teleost fish species’ swimming energetics is needed. Teleost fish in their natural environment show a great variety of locomotive strategies while feeding, migrating, and breeding [3]. Fish in aquaculture, however, are not subject to the same natural exposure in their swimming patterns. To translate these naturally evolved swimming capabilities into aquaculture, studies have characterized and described the swimming capabilities of fish. These swimming capabilities can be characterized into different stages: sustained swimming, where the fish can swim without fatigue, relying solely on aerobic red musculature; prolonged swimming, where fatigue occurs after a period of time, powered by both red musculature and some white glycolytic musculature; and burst swimming, which involves short, intense sprints that can only be maintained for brief periods, primarily powered by anaerobic white musculature [1,8,9].

Studies that have implemented flow regimes into RAS have shown improved growth and reduced stress in many salmonid species such as rainbow trout (Oncorhynchus mykiss), Arctic charr (Salvelinus alpinus), and Atlantic salmon (Salmo salar) [10,11,12,13]. Even for non-salmonid species, such as gilthead seabream (Sparus aurata) and yellowtail kingfish (Seriola lalandi) [14,15], sustained exercise has shown growth improvement and disease reduction. These studies have suggested that the water velocity implemented in sustained exercise regimes should be at a fish species’ optimal swimming speed (U_opt). Optimal swimming speed is the speed at which the metabolic rate is lowest per distance covered. Optimal swimming speed can also be described as the minimal cost of transport (COT_min), which is the least amount of energy used to move a given mass a certain distance [16], whereas preferred swimming speed (U_pref) refers to the speed at which fish choose to swim during voluntary activities, reflecting their performance for routine behaviors. It is influenced by factors such as environmental conditions and metabolic rate, and is typically a small speed range in both flowing and still water. U_pref is an important measure of fish behavior, energy efficiency, and ecological performance, and is used in various studies, including those related to aquaculture and migration [9].

The present study examined the relationship between preferred swimming speed (U_pref) and optimal swimming speed (U_opt) in rainbow trout at different temperatures, with a focus on non-migratory conditions typical of aquaculture environments. Specifically, the present study investigated if the optimal swimming speed (U_opt) reflects the preferred swimming speed (U_pref) in rainbow trout. The study aimed to determine rainbow trout U_opt using swim tunnel respirometry [17] and further determine the U_pref. Rainbow trout was used as a model species, due to its extensive appearance in fish physiology literature. The three experimental temperatures (10, 15, and 20 °C) were chosen, as they can each represent the lower, optimal, and upper thermal thresholds when studying the physiological performance of rainbow trout, and this is also consistent with previous research on the species [18]. This study demonstrates a new approach by implementing a circular raceway with consistent flow, different from other studies [9,19] that investigated volitional swimming.

2. Materials and Methods

2.1. Experimental Design

This study investigated the relationship between preferred swimming speed (U_pref) and optimal swimming speed (U_opt) in rainbow trout under different temperature conditions (10, 15, and 20 °C). All experiments agreed with the EU Directive 2010/63/EU for animal experiments and were performed with permission from the Danish Animal Experiments Inspectorate (License number: 2018-15-0201-01466). Two experimental setups were used: a circular raceway for volitional swimming (Figure 1) and a swim tunnel respirometer for forced-swimming tests (Figure 2). The experiments aimed to measure U_pref in a minimally invasive environment and compare it with U_opt, determined through a traditional forced-swimming respirometry method. The temperatures of the groups were selected as 20 for higher than optimal temperature and two temperatures lower than this. Holding and experimenting on the species, being easier at the lab below the optimum than above it, as the system water was 10. The specific values are represented in other literature as well, enabling comparison to other metrics.

2.2. Fish Husbandry

Rainbow trout, acquired from FREA Solutions, (n = 31, body length (fork), 24.5 ± 0.9 cm, wet mass 175.5 ± 20.1 g, $\overline{\mathrm{x}}$ ± SD) (

Table 1. Fish size measurements of the three treatment groups.

Temperature Acclimation Group	10 °C	15 °C	20 °C
n	10	10	11
Body length (cm)	24.2 ± 1.1	24.7 ± 0.5	24.7 ± 0.9
Mass (g)	187.5 ± 16.1	177.9 ± 12.7	162.7 ± 23.7
Height (cm)	6.0 ± 0.4	5.7 ± 0.5	5.1 ± 0.6
Width (cm)	2.7 ± 0.4	2.6 ± 0.2	2.6 ± 0.4

All measurements were made prior to trial initiation. The measurements are presented as $\overline{\mathrm{x}}$ ± SD. Height is the vertical/ventral-dorsal line, and width is the horizontal/lateral line when looking at the fish from the front.

), were housed in controlled lab conditions with the water temperature maintained at the experimental levels (10, 15, or 20 °C). Fish were fasted for 72 h prior to experiments to reduce postprandial effects. Salinity was kept at 30–33‰, and oxygen saturation was maintained at above 95%. All fish had been acclimated to seawater prior to experimentation.

2.3. Flow Calibration and Validation

The flow rate in the circular raceway (Figure 1) was mapped using a flow probe (Hoentzsch vane wheel) at 40 predetermined points across four transects (c.f. Figure 3). The flow rates at the 40 positions were used to make a 3D model and an equation in TableCurve3Dv4 describing the flow rate (z) at any position (x,y) of the tank and regression constants (a–g) (Equation (1)) (Illustrated in Figure 3 and Figure 4).

z = (a + b × x + c × y)/(1 + d × x + e × x2 + f × y + g × y2)

(1)

Similarly, water velocity in the swim tunnel was calibrated in Loligo Systems AutoResp. A flow probe (Hoentzsch vane wheel) was used to calibrate the motor voltage output to the accurate water velocity, and the solid blocking effect was corrected for each fish’s trial [20].

2.4. Preferred Swimming Speed (U_pref)

We report swimming speeds in body lengths per second (BL s⁻¹), where body length was determined as fork length. To measure U_pref, fish were introduced into a circular raceway system that allowed volitional swimming. The raceway consisted of a circular outer tank (diameter: 150 cm) and an inner circular perforated tank (diameter: 56 cm), creating a 47-centimeter-wide circular swimming section. The water depth in the swimming section was 20 cm. Consistent circular flow was achieved using a submersible pump (AL-KO Drain 7500 L h⁻¹ classic) in the inner tank that recirculated water to a vertical tube perforated at the submerged end, pointing tangentially along the wall at the edge of the outer tank wall (Figure 1). To maintain a constant temperature and steady supply of new water, three 500 W titanium rod heaters constantly heated the water. Aeration ensured oxygen saturation, and the supply of new seawater ensured that salinity was not increased by evaporation. Temperature regulation was achieved via a thermosensitive switch (Sonoff TH R3/Elite), which controlled a water pump (Eheim Universal Pump 1048) connected to an adjacent tank with 10 °C seawater, set to trigger ON when the water temperature exceeded a set threshold and OFF when it fell back below the respective experimental temperatures (10, 15, or 20 °C).

Fish were acclimated to the raceway for 18–20 h before data acquisition. A video camera mounted above the raceway continuously recorded fish position and movement for six hours at 25 frames per second, using commercial video surveillance software (H264 WebCam). Post-recording analysis was conducted using SwisTrack (Ver 4.2 2023, user guide available in Supplementary Materials) software to derive all x, y positions from the trials of each fish. Then, U_pref was determined using Equation (1).

The water flow at the location where the fish preferred to position themselves is circular in nature in the experimental setup, meaning that the fish are not swimming as if in a straight line. The curvature of the water flow causes an increase in the drag factor and forces the fish to increase its effort. The increased effort, corresponding to a higher swimming speed, can be accounted for by using the correction formula (Equation (2)) for swimming speed, as used in He and Wardle (1988) [21,22]. The correction factor for straight-line swimming is graphically illustrated in Figure 5.

U_s/U_v = [1 + (2(K_w/D_f^0.33)/C_d × L/R)² (D_f/D_w + K_m)²]^1/5.6

(2)

U_pref was determined by taking the median of the swimming speed values in BL s⁻¹, as distributions were multimodal. Note that the numbers obtained were median and IQR, c.f. Figure 6, average and standard deviation estimates, of the mode of highest frequency, which were completed by equating the mean of that distribution to the median and calculating SD for the mode as interquartile range divided by 1.35.

2.5. Optimal Swimming Speed (U_opt)

After completing the U_pref trials, fish were transferred to a swim tunnel respirometer (31.5 L Steffensen type) to measure swimming energetics (Figure 2; 3D model available at: a360.co/4g3OsFW). The fish were introduced to the swim tunnel at a swimming speed of 0.5 BL s⁻¹ for 12–16 h prior to the swimming test, in order to ensure acclimation to the new environment and reaching a baseline oxygen consumption (MO₂). Following the acclimation period, oxygen consumption rate was measured with gradually increasing swimming speeds by 0.5 BL s⁻¹ increments every three measurement cycles. The MO₂ was determined by 10 min intervals consisting of 240 s flush, 60 s wait, and 300 s measure-periods. Measurements were executed at 0.5, 1.0, 1.5, 2.0, and 2.5 BL s⁻¹ for each fish.

Trials continued until fish reached fatigue, defined as touching the back grid of the tunnel three consecutive times or ceasing forward motion. The relationship between swimming speed and MO₂ was fitted to the following exponential function:

MO₂ = a × e ^{(b × U)}

(3)

where a and b are constants, and U is fish swimming speed. U_opt was calculated as the speed with the least cost of transport (COT):

U_opt = b⁻¹

(4)

Critical swimming speed (U_pref) was determined as in Steinhausen et al. (2005). Lastly, standard metabolic rate (SMR), which is the minimal energy expenditure required for self-maintenance, was found by extrapolating Equation (3) back to x = 0 [23].

2.6. Data Analysis

Normality and variance homogeneity of the observed data of U_pref, U_opt, U_crit, and standard metabolic rate in each thermal treatment group (10, 15, and 20 °C) was determined by performing a Shapiro–Wilks test (α = 0.05). Once all necessary conditions were fulfilled, a statistical test was conducted on the following data.

A one-way ANOVA was conducted to assess any statistical differences (α = 0.05) between temperature treatment groups U_pref, U_opt, and U_crit, followed by post-hoc comparisons (Tukey’s honestly significant difference test (α = 0.05)).

The difference between U_pref and U_opt in the same treatment group (10, 15, and 20 °C) was determined by a paired t-test (α = 0.05), to assure all requirements were fulfilled; a Levene’s F-test was performed to test for equality in variance beforehand.

3. Results

3.1. Preferred Swimming Speed

After the introduction to the swimming section in the circular raceway, the fish spent 10–60 min exploring the new environment and then settled into a preferred position. The fish carried on in the same position for most of the measuring period. U_pref was determined for each thermal acclimation treatment group (10, 15, and 20 °C) (Figure 6), averaging volitional preferred swimming speeds of 1.18 ± 0.14, 1.17 ± 0.19, and 1.24 ± 0.15 BL s⁻¹, $\overline{\mathrm{x}}$ ± SD, respectively (

Table 2. Results from the preferred swimming speed and the swimming respirometer experiments of the three treatment groups.

Temperature Acclimation Group	10 °C	15 °C		20 °C
Upref (BL s−1)	1.2 ± 0.14	1.2 ± 0.19		1.2 ± 0.15
Uopt (BL s−1)	1.4 ± 0.19	1.5 ± 0.15		1.6 ± 0.24
Ucrit (BL s−1)	2.1 ± 0.35	2.6 ± 0.05		2.4 ± 0.42
MO2 @ 0.5 BL s−1 (mg O2 kg−1 h−1)	62.3 ± 4.8	99.2 ± 12.9		195.2 ± 30.9
AMR (mg O2 kg−1 h−1)	188.9 ± 44.6	386.0 ± 72.6		550.6 ± 171.2
SMR (mg O2 kg−1 h−1)	44.7	65.7		130.9
Aerobic scope @ 0.5 BL s−1	3.0 ± 0.6	3.8 ± 1.1		2.8 ± 0.7
Q10	2.5		3.8
	3.1

The results are presented as ($\overline{\mathrm{x}}$ ± SD), except for SMR, which is extrapolated from the average MO₂ measurements from swimming respirometer experiment. Q₁₀ is given between 10 and 15 °C, between 15 and 20 °C on the first row, and between 10 and 20 °C on the second row.

).

Statistical analysis showed no significant effect of temperature in preferred swimming speed between treatment groups (one-way ANOVA, p = 0.62). As the statistical test resulted in a large p-value, showing little to no variance in between the treatment groups, no further analysis was performed.

3.2. Energetics and Optimal Swimming Speed

The fish spent 10–30 min settling into the new environment and began swimming steadily at 0.5 BL s⁻¹. After 1–2 h, the fish reached a steady oxygen consumption (MO₂: mg O₂ kg⁻¹ h⁻¹). The MO₂ @ 0.5 BL s⁻¹ of each treatment group (10, 15, and 20 °C) was determined post-experiment, and the average of each group was 62.3 ± 4.9, 99.2 ± 12.5, and 195.2 ± 30.9 mg O₂ kg⁻¹ h⁻¹, $\overline{\mathrm{x}}$ ± SD, respectively. From the MO₂ measurements, the average SMR was found by extrapolating back to 0 BL s⁻¹ for each treatment group; 44.70, 65.70, and 130.90 mg O₂ kg⁻¹ h⁻¹, respectively. Active metabolic rate (AMR) was determined as the highest MO₂ measurement from the swimming respirometer trial; 188.9 ± 44, 386.0 ± 72.7, and 550.6 ± 171.2 mg O₂ kg⁻¹ h⁻¹, $\overline{\mathrm{x}}$ ± SD, respectively.

Lastly, the Fractional Aerobic Scope (AS) was determined for each treatment group (10, 15, and 20 °C), as AMR divided by SMR; AS: 3.0 ± 0.6,3.8 ± 1.1, and 2.8 ± 0.7, $\overline{\mathrm{x}}$ ± SD, respectively. The Q₁₀ values between the treatment groups of the determined MO₂ @ 0.5 BL s⁻¹ were Q₁₀ = 2.5 between 10 °C and 15 °C, 3.1 between 15 °C and 20 °C, and 3.8 between 10 °C and 20 °C. Statistical analysis showed a significant difference between the MO₂ @ 0.5 BL s⁻¹ in the three groups (one-way ANOVA, p = 0.055). A further post-hoc (Tukey’s HSD) test showed significant differences among all groups.

The U_opt for each treatment group (10, 15, and 20 °C) was 1.4 ± 0.19, 1.5 ± 0.15, and 1.6 ± 0.24 BL s⁻¹, $\overline{\mathrm{x}}$ ± SD, respectively. The statistical analysis showed no significant difference between the three groups (one-way ANOVA, p = 0.057). As the p-value was only slightly above the significance level, a further post-hoc (Tukey’s HSD) showed a significant difference between the U_opt at 10 °C and 20 °C (p = 0.046).

Furthermore, the U_crit results for each treatment group (10, 15, and 20 °C) were 2.1 ± 0.35, 2.6 ± 0.05, and 2.4 ± 0.42 BL s⁻¹, $\overline{\mathrm{x}}$ ± SD, respectively. Statistical analysis showed significant differences between groups (one-way ANOVA, p = 0.19). A further post-hoc (Tukey’s HSD) test showed a significant difference between treatment groups 10 °C and 15 °C, and treatment groups 10 °C and 20 °C (p = 0.009 and p = 0.049, respectively). All results from the swimming respirometry trials are presented in Table 2 and graphically illustrated in Figure 7.

3.3. Cost of Transport

The average COT_min for each temperature (10, 15, and 20 °C) was 83.4 ± 9.3, 118.8 ± 14.7, and 210.7 ± 23.4 mg O₂ kg⁻¹ BL s⁻¹, $\overline{\mathrm{x}}$ ± SD, respectively. The statistical test revealed a significant difference in COT_min between thermal groups (one-way ANOVA, p = 1.781 × 10⁻¹³. A further post-hoc (Tukey’s HSD) test revealed a significant difference among all groups (Figure 7, lower).

3.4. Comparison of Swimming Speeds

Homogeneity across all groups of the U_pref and U_opt experiments was tested and showed no significant difference in variance (Levene’s, F = 1.13, p = 0.3576). A significant difference between U_pref and U_opt across all treatment groups (T: −4.260, p = 0.00027) was determined (Figure 8).

The oxygen consumption swimming at U_pref was 18.9, 20.4, and 22.1% lower than swimming at U_opt at 10, 15, and 20 °C, respectively.

4. Discussion

The present study aimed to deepen our understanding of rainbow trout swimming energetics by examining how their preferred swimming speed (U_pref) relates to the optimal swimming speed (U_opt). Prior efforts often relied on forced-swimming trials to determine U_opt for aquaculture conditions, assuming that fish would naturally prefer this energetically optimal speed. Our findings challenge this assumption. Despite changes in metabolic rates and aerobic scope with temperature, the chosen U_pref remained near 1.2 BL s⁻¹ across all tested temperatures (10, 15, and 20 °C), while U_opt was consistently and significantly higher (1.4–1.6 BL s⁻¹). This discrepancy suggests that rainbow trout should be swimming at U_pref rather than U_opt in aquaculture environments. In addition, swimming at U_pref is 18–22% less energetically costly than swimming at U_opt.

These results differ from the conclusions on a study on brook charr Salvelinus fontinalis [9], where preferred and optimal speeds were identical at 15 °C, but the authors reported large variance in the results. It appears from the distribution of U_pref that a larger amount of time was spent at the two categories below U_opt than at the two at U_opt and just above ([9] Figure 3). Thus, using a different analytical approach by Tudorache et al. [9] could have led to the same conclusion as ours, that U_pref is below U_opt.

This conclusion is in contrast to the conventional thinking in terms of minimizing cost per distance traveled. Our cost of transport (COT) analysis further shows that small deviations from U_opt (e.g., at 10 and 15 °C) do not drastically alter the cost of transport, making it less critical for fish to fine-tune their speed to a strict physiologically forced optimum.

Metabolic rates (SMR, AMR) and aerobic scope were temperature dependent. SMR increased with temperature, yielding Q₁₀ values (2.5–3.8) similar to those found elsewhere [24,25]. Aerobic scope was largest at 15 °C, aligning with the known temperature range (15.8–17 °C) for rainbow trout and with the highest U_crit observed [26,27]. This matches previous findings that maximal aerobic capacity occurs around this temperature [28]. Similar ranges in U_opt (1–2 BL s⁻¹) have been reported for other salmonid species [29,30], and ~1.5 BL s⁻¹ has previously been recommended for growth in aquaculture. However, due to spontaneous locomotor activity, actual swimming speeds vary widely from 0.5 to 3.0 BL s⁻¹ [31,32,33].

Conventional forced-swimming methods significantly constrain fish movement, limiting their ability to exhibit natural swimming behaviors. In contrast, the circular raceway used in this study provides a more natural and less restrictive environment. While boundary effects on fish swimming may exist in the raceway, these effects are minimized by its continuous design, allowing the fish to swim without encountering abrupt barriers. This setup better allows for spontaneous locomotor behavior and enables studying effects on fishes using a more top–down perspective.

Lastly, the comparison between the results from the preferred swimming speed (U_pref) and the forced swimming speed (U_opt) showcase that in aquaculture settings, rainbow trout are more inclined to swim at U_pref instead of U_opt, as swimming at U_pref is 18–22% more energy efficient. The results also suggest that minor variations from U_opt do not significantly impact the cost of transport, suggesting that strict adherence to U_opt is not crucial for energy efficiency.

5. Conclusions

Unlike forced-swimming tests, this method for determining U_pref was less invasive, allowing fish to swim voluntarily in a circular raceway, mimicking aquaculture conditions.

Temperature significantly influenced metabolic rates and aerobic scope, with SMR and AMR increasing at higher temperatures, yet U_pref remained consistent (~1.2 BL s⁻¹) across 10, 15, and 20 °C. U_opt varied slightly (1.4–1.6 BL s⁻¹) and was consistently higher than U_pref. Swimming at U_pref reduced energy costs by 18–22% compared to U_opt, making it an advantageous choice for aquaculture systems aiming to improve growth efficiency and fish welfare.