Interpopulation variation in growth, CTMax and metabolism among seasonal phenologies of Chinook Salmon

Fish underwent metabolic trials in one of four, 5 L automated swim tunnel respirometers (Loligo, Denmark). The four tunnels were split into two paired systems with two tunnels sharing a single sump and heat pump. Water for each swim tunnel system was pumped (PM700, Danner USA) from the sump into an aerated water bath surrounding each swim tunnel, and then returned to the sump. Sumps were supplied with non-chlorinated fresh water from a designated well and aerated with air stones. The temperature of the sump (and therefore the swim tunnels) was maintained (±0.5°C) by circulating water through a heat pump (model DSHP-7; Aqua Logic Delta Star, USA) using a high-volume water pump (Sweetwater SHE 1.7 Aquatic Ecosystems, USA). In addition, each sump contained a thermostatically controlled titanium heater (TH-800; Finnex, USA). Swim tunnels and associated sump systems were cleaned and sanitized with bleach weekly to reduce potential for bacterial growth. Dissolved oxygen saturation within the swim tunnels was measured using fibre-optic dipping probes (Loligo OX11250) which continuously recorded data via AutoResp™ software (version 2.3.0). Oxygen probes were calibrated weekly using a two-point, temperature-paired calibration method. Water velocity of the swim tunnels was quantified and calibrated using a flowmeter (Hontzcsh, Germany) and regulated using a variable frequency drive controller (models 4x and 12K; SEW Eurodrive, USA). The velocity (precision <1 cm s-1) for each tunnel was controlled remotely using the Autoresp™ program and a DAQ-M data acquisition device (Loligo, Denmark). Swim tunnels were surrounded by shade cloth to reduce disturbance of the fish. Fish were remotely and individually monitored using infrared cameras (QSC1352W; Q-see, China) connected to a computer monitor and DVR recorder. Oxygen consumption rates for both routine and maximum metabolic rates were captured using intermittent respirometry(Brett 1964). Flush pumps (Eheim 1048A, Germany) for each tunnel pumped aerated fresh water through the swim chamber and was automatically controlled via the AutoResp™ software and DAQ-M system. This system would seal the tunnel and enable the measurement of oxygen consumption attributable to the fish. Oxygen saturation levels were not allowed to drop below 80% and restored within three minutes once the flush pump was activated. Oxygen saturation data from AutoResp™ was transformed to oxygen concentration using the following equation: Where %O2Sat is the oxygen saturation percentage reported from AutoResp™; αO2 is the coefficient temperature-corrected oxygen solubility (mgO2 L-1 mmHg-1); and BP is the barometric pressure (mmHg). Oxygen concentration (milligrams of oxygen per liter) was measured every second and regressed over time, the coefficient of this relationship (milligrams of oxygen per liter per second) was then converted to metabolic rate (milligrams of oxygen per kilogram per minute, Equation 3). Where R is the calculated coefficient of oxygen over time; V is the volume of the closed respirometer; M is the mass of the fish in kilograms and ’60’ transforms the rate from per second to per minute. An allometric scaling exponent was not incorporated due to similarity in fish sizes and to maximize comparability with metabolic data from the Mokelumne Hatchery (CA) fall-run population (Poletto et al. 2017). Routine Metabolic Rate Prior to routine metabolic rate (RMR) trials fish were fasted to ensure a post-prandial state. Fish reared at 16 or 20°C were fasted for 24 hours, while fish acclimated to 11°C were fasted for 48 hours. Fish were then transferred into a swim tunnel respirometer between 13:00 and 17:00. After a 30-minutes at their acclimation temperature the temperature was adjusted at 2°C h-1 to the test temperature (8 – 26°C). Automated intermittent flow respirometry began 30 minutes after the test temperature was achieved and continued overnight. Measurement periods ranged from 900 to 1800 seconds in duration, flush periods were 180-300 seconds. Periods varied in length in response to fish size and test temperature to ensure oxygen saturation was kept high (>80%) during the trial. A small circulation pump (DC30A-1230, Shenzhen Zhongke, China) ensured that water was mixed without disturbing the fish. Fish activity was monitored by overhead infra-red cameras and measurement periods when the fish were active were discarded. RMR was calculated by averaging the three lowest RMR values(Poletto et al. 2017). RMR measurements were concluded by 08:00 ± 40 min. Maximum Metabolic Rate A modified critical swimming velocity protocol was used to elicit maximal metabolic rate (MMR)(Poletto et al. 2017). Tunnel speed was increased gradually from 0 to 30 cm s-1 over an ~2 min period and held there for 20 min. For each subsequent 20-min measurement period, tunnel velocity was increased 10% up to a maximum of 6 cm s-1 per step. Fish were swum until exhausted and unable to swim. Swimming metabolism was measured by sealing the tunnel for approximately 16 minutes of the 20-minute measurement period. When a fish became impinged upon the back screen (>2/3 of body in contact with screen) the tunnel velocity was stopped for ~1 minute and then gradually returned to the original speed over 2 minutes. A fish was determined to be exhausted if it became impinged twice within the same velocity step. At this point the tunnel impellor was stopped to allow for recovery. The highest metabolic rate measured over a minimum of 5 minutes during active swimming was taken as the MMR. Post-experiment, the tunnel was returned to the acclimation temperature and fish were transferred to a recovery tank and monitored. In seeking evidence of metabolic collapse at near-critical temperatures, some metabolic trials were conducted at temperatures exceeding the tolerance of the fish. These mortality events represent potential lethal upper limits for sub-acute thermal persistence (Fig. S1). Data from fish which did not survive the trial or recovery were not used in analysis. After a 24-hour recovery period fish were euthanized in a buffered solution of MS-222 (0.5g/L). Measurements for mass (g), fork length (cm) and total length (cm) were taken, and Fulton’s condition factor was calculated. Aerobic scope (AS) was calculated as the difference between a fish’s RMR and MMR. Thermal optimums (TOPT) were defined as the temperature when aerobic scope was maximized, and calculated as the root-value of the derivative of the quadratic function describing the relationship between AS and test temperature. Growth Data Growth measurements were initiated in mid to late spring when all populations would still be rearing prior to outmigration. Growth data were gathered every two weeks by measuring a sample of 30 fish from each treatment (n=15 per tank, n = 1528 total measurements). Fish were not individually marked and therefore growth rate was calculated across individuals. Fish were arbitrarily netted from their treatment tank and transferred to an aerated five-gallon bucket until measured. Fish were air exposed for ~15-20 seconds to measure mass (± 0.01 grams, Ohaus B3000D) and fork length (± 0.1 cm) and then placed into a second bucket for recovery before returning to their original treatment tank. Fish were netted and measured by the same experimenter across all sampling days. Condition factor was calculated as Fulton's condition factor (K) using the equation K = 100*Mass/Fork Length^3. Critical Thermal Maxima Critical Thermal Maximum (CTMax) values were quantified according to established methods, briefly described below1. We placed six 4L Pyrex beakers in a fiberglass bath tray (1m x 2m x .2m). Beakers were aerated with an air stone to ensure both adequate oxygen saturation and circulation of water within the beaker. The volume of water in each individual beaker (approx. 2.5 L) was calibrated to ensure even heating across all CTMax beakers (0.33°C/min). Two pumps (PM700, Danner USA) were used to circulate water: one pump recirculated water across three heaters (Process Technology S4229/P11), while the other distributed heated water through the CTMax bath via a distribution manifold. Experiments began with water temperature set at the fish’s acclimation temperatures (11, 16 or 20°C). Fish of appropriate size (n = 377, 12.4 ± 0.83 cm) were arbitrarily selected from treatment tanks and transferred to separate tanks for fasting. To ensure fish were in a similar postprandial state, fish reared at 20°C and 16°C were fasted for 24 hours and 11°C fish were fasted for 48 hours to account for their slower metabolic rate. Once fasted, fish were individually netted and transferred into individual beakers within the CTMax heat bath. Fish were given 30 minutes to acclimate to their CTMax beaker after which the CTMax trial began. During the CTMax trial, beaker temperature was taken every 5 minutes using a thermocouple (Omega HH81A). Thermocouple measurements were calibrated to a Fisherbrand® NIST certified mercury thermometer following each trial. Fish were observed continually for signs of distress and loss of equilibrium. The CTMax trial endpoint was loss of equilibrium, at which point the temperature of the CTMax beaker was recorded2,3. Fish were then removed and retuned to a recovery bath at their acclimation temperature. Fish that did not fully recover within 24-hours were not included in analysis (6% of individuals). After the 24-hr recovery, fish were weighed (wet mass ± 0.01g) and measured (fork length ± 0.1 cm). N ≥ 20 for all treatments (population x. acclimation temperature) except for winter-run fish reared at 16 (n= 17) or 20°C (n=9). Winter-run reared at 20°C were limited due to a mortality event. On October 17th 2018, an outbreak of Columnaris in a single tank of winter-run Chinook salmon rearing at 20°C resulted in the mortality of the remaining tank population (n=7). Necropsy of the salmon indicated empty stomachs. The mortality of this population is hypothesized to be a result of thermal stressed after being reared at 20°C for so long (202 days). CTM data for the 20°C acclimation group was limited to 9 fish tested 41 days prior to the mortality event. Missing values: Only fish that successfully completed the Critical Thermal Maximum (CTMax) trials are included in this dataset. Fish that did not recover from the trial were excluded from analysis, and therefore this dataset. The CTMax trial is designed to be survived and therefore failure to recover fully is reflective of an abnormality in the fish’s physiology or experimenter error, both of which could bias the ultimate results. 1. Becker, C. D. & Genoway, R. G. Evaluation of the critical thermal maximum for determining thermal tolerance of freshwater fish. Environmental Biology of Fishes 4, 245–256 (1979). 2. Beitinger, T. L., Bennett, W. A. & McCauley, R. W. Temperature tolerances of North American freshwater fishes exposed to dynamic changes in temperature. Environmental Biology of Fishes 58, 237–275 (2000). 3. Fangue, N. A., Hofmeister, M. & Schulte, P. M. Intraspecific variation in thermal tolerance and heat shock protein gene expression in common killifish, Fundulus heteroclitus. Journal of Experimental Biology 209, 2859–2872 (2006).

实验鱼类在4台5L容量的自动游泳隧道式呼吸仪（automated swim tunnel respirometer，Loligo，丹麦）中开展代谢试验。4台呼吸仪分为两套配对系统，每台系统包含2台呼吸仪，共享同一个集液槽（sump）与热泵。每套游泳隧道系统的供水由水泵（PM700，Danner，美国）从集液槽抽取，注入环绕每台呼吸仪的曝气水浴槽，之后流回集液槽。集液槽由指定水井供给不含氯的淡水，并通过气石曝气。集液槽（及呼吸仪）的水温通过热泵（DSHP-7型，Aqua Logic Delta Star，美国）搭配大流量水泵（Sweetwater SHE 1.7 Aquatic Ecosystems，美国）循环水体维持，控温精度为±0.5℃。此外，每套集液槽均配备一台恒温控制钛加热器（TH-800，Finnex，美国）。

每周需对呼吸仪及配套集液槽系统进行漂白清洁消毒，以降低细菌滋生风险。

游泳隧道内的溶解氧饱和度通过光纤浸入式探头（Loligo OX11250）实时监测，数据通过AutoResp™软件（版本2.3.0）持续记录。氧探头每周采用两点温度配对校准法进行校准。

游泳隧道的水流速度采用流量计（Hontzcsh，德国）定量校准，并通过变频驱动控制器（4x及12K型，SEW Eurodrive，美国）调节。每台呼吸仪的流速（精度<1 cm·s⁻¹）通过Autoresp™软件与DAQ-M数据采集装置（Loligo，丹麦）远程控制。

游泳隧道外包裹遮光布，以减少对鱼类的干扰。鱼类通过连接电脑显示器与硬盘录像机（DVR）的红外摄像头（QSC1352W，Q-see，中国）实现远程单独监测。

采用间断式呼吸计量法（Brett 1964）测定鱼类的日常代谢率（routine metabolic rate, RMR）与最大代谢率（maximum metabolic rate, MMR）。每台呼吸仪配备冲水水泵（Eheim 1048A，德国），可将曝气淡水泵入呼吸腔，该操作由AutoResp™软件与DAQ-M系统自动控制。系统会密封呼吸隧道，以测定鱼类引起的耗氧量。

实验过程中溶解氧饱和度不得低于80%，启动冲水泵后需在3分钟内恢复至要求水平。AutoResp™软件记录的氧饱和度数据通过以下公式转换为氧浓度：其中%O₂Sat为AutoResp™报告的氧饱和度百分比；αO₂为经温度校正的氧溶解系数（mgO₂·L⁻¹·mmHg⁻¹）；BP为大气压（mmHg）。

氧浓度（单位：毫克氧每升）每秒采集一次，并随时间进行线性回归，所得回归系数（单位：毫克氧每升每秒）随后转换为代谢率（单位：毫克氧每千克每分钟，公式3）。其中R为氧浓度随时间变化的回归系数；V为密闭呼吸仪的体积；M为鱼类体重（千克）；数字‘60’用于将速率单位从每秒转换为每分钟。

由于受试鱼类体型相近，且为最大化与Mokelumne孵化场（加利福尼亚州）秋季洄游种群代谢数据的可比性（Poletto等，2017），本实验未引入异速生长缩放指数。

### 日常代谢率
日常代谢率（RMR）试验前，鱼类需禁食以确保处于餐后空腹状态。饲养于16℃或20℃的鱼类禁食24小时，而适应11℃的鱼类禁食48小时。随后于13:00至17:00之间将鱼类转移至游泳隧道式呼吸仪。在适应温度下静置30分钟后，以2℃·h⁻¹的速率将水温调节至试验温度（8~26℃）。达到试验温度30分钟后启动自动间断式水流呼吸计量，试验持续整夜。测量时段时长为900~1800秒，冲水时段为180~300秒。时段长度根据鱼类体型与试验温度调整，以确保试验过程中溶解氧饱和度维持在较高水平（>80%）。小型循环水泵（DC30A-1230，深圳中科，中国）可确保水体充分混合且不干扰鱼类。

通过顶部红外摄像头监测鱼类活动，若鱼类处于活跃状态，则舍弃对应时段的测量数据。RMR通过取三项最低RMR值的平均值计算得到（Poletto等，2017）。RMR测量需在08:00±40分钟内完成。

### 最大代谢率
采用改良临界游泳速度方案诱导鱼类达到最大代谢率（MMR）（Poletto等，2017）。隧道流速在约2分钟内从0逐步提升至30 cm·s⁻¹，随后维持该流速20分钟。后续每个20分钟测量时段，流速提升10%，每步最大增幅为6 cm·s⁻¹。试验持续至鱼类力竭无法游泳为止。

在20分钟的测量时段内，密封呼吸隧道约16分钟以测定游泳代谢率。当鱼类贴附于后挡板（超过2/3躯体与挡板接触）时，暂停隧道流速约1分钟，随后在2分钟内逐步恢复至原流速。若鱼类在同一流速步幅内两次贴附挡板，则判定其力竭。此时停止隧道叶轮，使鱼类恢复。

将鱼类活跃游泳期间至少5分钟内测得的最高代谢率作为MMR。试验结束后，将隧道水温恢复至适应温度，将鱼类转移至恢复箱并监测状态。

为探究近临界温度下的代谢衰竭现象，部分代谢试验在超过鱼类耐受上限的温度下开展。此类死亡事件代表亚急性热耐受的潜在致死上限（补充图S1）。未在试验或恢复阶段存活的鱼类数据未纳入分析。

经过24小时恢复后，使用缓冲MS-222溶液（0.5g/L）对鱼类实施安乐死。测量其体重（克）、叉长（cm）与全长（cm），并计算富尔顿条件系数。有氧代谢范围（aerobic scope）为鱼类日常代谢率与最大代谢率的差值。热最适温度（thermal optimums, TOPT）定义为有氧代谢范围达到最大值时的温度，通过描述有氧代谢范围与试验温度关系的二次函数导数的根值计算得到。

### 生长数据
生长测量于春季中晚期启动，此时所有种群仍处于洄游前的饲养阶段。每两周采集一次生长数据，从每个处理组中取样30尾鱼类（每个养殖箱15尾，总计1528次测量）。由于未对鱼类进行个体标记，因此生长速率基于群体个体计算。

随机用网从处理养殖箱中捞取鱼类，转移至曝气的五加仑水桶中等待测量。鱼类暴露于空气约15~20秒以测量体重（精度±0.01克，Ohaus B3000D）与叉长（精度±0.1 cm），随后转移至第二个水桶中恢复，之后放回原处理养殖箱。所有采样日均由同一实验人员捞取并测量鱼类。条件系数采用富尔顿条件系数（K）计算，公式为K = 100×体重/叉长³。

### 临界热最大值
临界热最大值（Critical Thermal Maximum, CTMax）的测定参照已发表方法，简述如下¹。我们将6个4L Pyrex烧杯置于玻璃钢浴槽（1m×2m×0.2m）中。通过气石对烧杯曝气，确保烧杯内水体具有充足的溶解氧饱和度与良好的循环。每个烧杯的水体体积（约2.5L）经过校准，以确保所有CTMax烧杯的升温速率均匀（0.33℃/min）。

采用两台水泵（PM700，Danner，美国）循环水体：一台水泵将水循环通过三台加热器（Process Technology S4229/P11），另一台水泵通过分配歧管将加热后的水循环至CTMax浴槽。

试验起始水温设置为鱼类的适应温度（11、16或20℃）。从处理养殖箱中随机选取规格合适的鱼类（n=377，12.4±0.83 cm），转移至单独暂养箱中禁食。为确保所有鱼类处于相似的空腹状态，饲养于20℃与16℃的鱼类禁食24小时，饲养于11℃的鱼类禁食48小时，以适配其较低的代谢速率。

禁食完成后，将鱼类单独捞取并转移至CTMax热浴槽内的单个烧杯中。鱼类在烧杯中适应30分钟后，启动CTMax试验。试验期间，每5分钟使用热电偶（Omega HH81A）测量一次烧杯水温。每次试验结束后，将热电偶测量值与Fisherbrand® NIST认证水银温度计进行校准。

持续观察鱼类是否出现应激与平衡丧失的迹象。CTMax试验的终点为平衡丧失，此时记录烧杯内水温¹,²,³。随后将鱼类取出，转移至适应温度的恢复浴槽中。若鱼类在24小时内未完全恢复，则不纳入分析（占个体总数的6%）。

24小时恢复后，称量鱼类体重（湿重±0.01g）并测量叉长（±0.1 cm）。所有处理组（种群×适应温度）的样本量N≥20，仅16℃饲养的冬季洄游种群（n=17）与20℃饲养的冬季洄游种群（n=9）除外。

20℃饲养的冬季洄游奇努克鲑鱼（Chinook salmon）种群样本量受限源于一次死亡事件：2018年10月17日，单缸饲养于20℃的该种群爆发柱状黄杆菌病（Columnaris），导致该缸剩余种群全部死亡（n=7）。解剖检查显示鱼类胃部空虚。推测该种群的死亡源于长期饲养于20℃（202天）引发的热应激。20℃适应组的CTM数据仅包含死亡事件发生前41天测试的9尾鱼类。

### 缺失值说明
本数据集仅纳入成功完成临界热最大值（CTMax）试验的鱼类。未从试验中恢复的鱼类被排除在分析与本数据集之外。CTMax试验设计为可存活的，因此未能完全恢复的情况反映了鱼类生理异常或实验人员操作失误，二者均可能偏倚最终结果。

### 参考文献
1. Becker, C. D. & Genoway, R. G. 用于测定淡水鱼类热耐受的临界热最大值评估方法. 鱼类生态学与环境生物学 4, 245–256 (1979).
2. Beitinger, T. L., Bennett, W. A. & McCauley, R. W. 暴露于动态温度变化的北美淡水鱼类的温度耐受范围. 鱼类生态学与环境生物学 58, 237–275 (2000).
3. Fangue, N. A., Hofmeister, M. & Schulte, P. M. 普通鳉（Fundulus heteroclitus）种内热耐受与热休克蛋白基因表达的变异. 实验生物学杂志 209, 2859–2872 (2006).