Effects of incubation temperature on the upper thermal tolerance of the imperiled longfin smelt (Spirinchus thaleichthys)

Abstract

Upper thermal limits in many fish species are limited, in part, by the heart's ability to meet increased oxygen demand during high temperatures. Cardiac plasticity induced by developmental temperatures can therefore influence thermal tolerance. Here, we determined how incubation temperatures during the embryonic stage influence cardiac performance across temperatures during the sensitive larval stage of the imperiled longfin smelt. We transposed a cardiac assay for larger fish to newly hatched larvae that were incubated at 9°C, 12°C or 15°C. We measured heart rate over increases in temperature to identify the Arrhenius breakpoint temperature (T_AB), a proxy for thermal optimum and two upper thermal limit metrics: temperature when heart rate is maximized (T_peak) and when cardiac arrhythmia occurs (T_Arr). Higher incubation temperatures increased T_AB, T_peak and T_Arr, but high individual variation in all three metrics resulted in great overlap of individuals at T_AB, T_peak and T_Arr across temperatures. We found that the temperatures at which 10% of individuals reached T_peak or T_Arr and temperatures at which number of individuals at T_AB relative to T_peak (ΔN(T_AB,T_peak)) was maximal, correlated more closely with upper thermal limits and thermal optima inferred from previous studies, compared to the mean values of the three cardiac metrics of the present study. Higher incubation temperatures increased the 10% T_peak and T_Arr thresholds but maximum ΔN(T_AB,T_peak) largely remained the same, suggesting that incubation temperatures modulate upper thermal limits but not T_opt for a group of larvae. Overall, by measuring cardiac performance across temperatures, we defined upper thermal limits (10% thresholds; T_peak, 14.4-17.5°C; T_Arr, 16.9-20.2°C) and optima (ΔN(T_AB,T_peak), 12.4-14.4°C) that can guide conservation strategies for longfin smelt and demonstrated the potential of this cardiac assay for informing conservation plans for the early life stages of fish.

Keywords: Arrhenius breakpoint; San Francisco Estuary; cardiac performance; climate change; heart rate; larvae; thermal performance.

Figure 1

The effect of increasing temperature on heart rate (f_H) of longfin smelt larvae. Representative plot of heart rate with temperature (A) and Arrhenius plot of f_H with increasing temperatures (B) for an individual larva. Average heart rate with temperature for all individuals (C) and Arrhenius plots of f_H with increasing temperatures (D) for different incubation temperature groups. Regression lines for representative Arrhenius plots are divided into two segments, representing the method by which Arrhenius breakpoint temperatures (T_AB) were determined for each individual. Unlike individual Arrhenius plots, grouped Arrhenius plots were best fit as 3 segments due to variation in post-T_AB changes in f_H across individuals. Dotted line indicates temperature where heart rate peaks (T_peak), dash-dotted line indicates temperature where cardiac arrhythmia was first observed (T_Arr), and dashed line indicates the T_AB. Data presented as mean ± SEM.

Figure 2

Median cardiac metrics do not correlate with inferred optimal temperatures or upper thermal limits. Arrhenius breakpoint temperature (T_AB) (A), temperature where heart rate reaches a maximum (T_peak) (B), and temperature where arrhythmia first occurs (T_Arr) (C) of one day post-hatch longfin smelt larvae incubated at and held for one day at different temperatures. Boxplots represent median, first and third quartiles of cardiac performance metrics. Shaded dark gray area indicates the temperature range in which longfin smelt are cultured at the Fish Conservation and Culture Laboratory (Supplementary Table 3) and light gray shaded area indicates the temperatures where longfin smelt larvae are most abundant in the San Francisco Estuary. Dotted lines indicate the temperature beyond which longfin smelt larvae are no longer observed in the field and dashed lines indicate the temperature where reduced larval growth rates have been measured. Different letters indicate statistical significance between incubation temperatures within each cardiac performance metric.

Figure 3

The 10% threshold for T_peak and T_Arr correlate more closely with indicators of upper thermal limits measured by previous studies. Proportion of individuals reaching temperatures where heart rate reaches a maximum (T_peak) (A) and where arrhythmia first occurs (T_Arr) (B) as test temperatures increased. Embryos were incubated at 9°C, 12°C, and 15°C. Gray shaded areas indicate temperatures beyond those tested experimentally and are thus predictions of the models. Horizontal lines indicate 10, 50 and 95% thresholds and temperatures listed correspond to the temperature at which these thresholds were met for each incubation temperature. Dotted lines indicate the temperature beyond which longfin smelt larvae are no longer observed in the field and dashed lines indicate the temperature where reduced larval growth rates have been measured.

Figure 4

Overlays of proportion of individuals at Arrhenius breakpoint temperatures (T_AB) and temperatures where heart rate is maximized (T_peak) and experience arrhythmia (T_Arr) reveal a temperature range where the proportion of individuals at T_AB relative to T_peak or T_Arr is maximized. Overlay of the proportion of individuals at T_AB and that have reached T_peak (A-C) and T_Arr (D-F) for 9°C (A,D), 12°C (B,E), and 15°C (C,F) incubation temperature groups. Smooths of proportion of individuals at T_AB are kernel density estimates of proportion of individuals at T_AB. Dashed and dotted curves are logistic curves of the proportion of individuals at T_peak and T_Arr. Logistic curves are the same as in Fig. 3. Hashed areas under curve indicate the temperature ranges where T_AB-T_peak and T_AB-T_Arr are maximized. Dark gray shaded area indicates temperature range that longfin smelt larvae are cultured at the Fish Conservation and Culture Laboratory (Supplementary Table 3), and light gray shaded areas indicate temperatures where larvae are most abundant in the San Francisco Estuary. Dotted lines indicate the temperature beyond which longfin smelt larvae are no longer observed in the field and dashed lines indicate the temperature where reduced larval growth rates have been measured.

Figure 5

Smoothed fits (s(Temperature)) from GAMs of the relationship between number of individuals at T_AB minus those that have reached T_peak (∆N(T_AB,T_peak)) (A) and relationship between number of individuals at T_AB minus those that have reached T_Arr (∆N(T_AB,T_Arr)) (B) across temperatures for each incubation temperature. Vertical colored lines and associated temperatures indicate the temperature at which ∆N(T_AB,T_peak) and ∆N(T_AB,T_Arr) begin decreasing with temperature. Plots are fitted smooths and 95% confidence intervals from GAMs. The y-axis units are centered on zero. The estimated degrees of freedom of the smooths for each incubation temperature are: ∆N(T_AB,T_peak): 9°C: 4.46, 12°C: 4.49, 15°C: 3.45, ∆N(T_AB,T_Arr): 9°C: 5.61, 12°C: 5.18, 15°C: 4.49. Dark gray shaded area indicates temperature range that longfin smelt larvae are cultured at the Fish Conservation and Culture Laboratory (Supplementary Table 3), and light gray shaded areas indicate temperatures where larvae are most abundant in the San Francisco Estuary. Dotted lines indicate the temperature beyond which longfin smelt larvae are no longer observed in the field and dashed lines indicate the temperature where reduced larval growth rates have been measured.