Introduction

Many populations of shortnose sturgeon, Acipenser brevirostrum, are severely depleted throughout the sturgeon’s native range, from the Saint John River in New Brunswick, Canada to the St. Johns River, Florida (Vladykov and Greeley 1963). Commercial fishing in the late nineteenth century decimated many populations of shortnose sturgeon (Dadswell et al. 1984) and led to the species’ protection under the Endangered Species Act (Kynard 1997). However, the complex life history of shortnose sturgeon makes managing the species difficult; as a result, many populations remain severely depleted. Degradations of habitat and water quality are believed to contribute to reduced juvenile recruitment (Smith and Collins 1996; Collins et al. 2000); however, a lack of field studies has resulted in a poor understanding of habitat requirements for young-of-the-year (YOY) shortnose sturgeon (Kynard 1997).

Hydrologic alterations, such as the construction of dams, dredging, and water removal for waste treatment, agriculture, and drinking water all reduce river discharges and alter thermal regimes of rivers (Baxter 1977; Lydeard and Mayden 1995; Poff et al. 1997; Bunn and Arthington 2002; Pringle 2003). Thermal effluents, such as those released by power plants, cause further increases in water temperatures (Baxter 1977; Pringle 2003). River discharges may decline in summer periods of increased temperatures and reduced precipitation. As a result, summer water temperatures in coastal rivers of the southeastern United States may exceed 31°C for up to 2 months (July and August) during low-discharge years (Dyar and Alhadeff 1997; DeVries 2006). Such high water temperatures may be extremely stressful and potentially lethal to YOY shortnose sturgeon, which would be ≤200 days old and ≤200 mm TL at the time of exposure to these high temperatures (Ziegeweid 2006).

Studies that examine the effects of temperature on sturgeon populations are limited. High temperature and hypoxia have been shown to reduce growth rates, increase mortality, and reduce swimming activities of white sturgeon, Acipenser transmontanus (Cech et al. 1984; Crocker and Cech 1997). Respiration and survival rates decreased for juvenile Atlantic sturgeon, Acipenser oxyrinchus, exposed to hypoxia and high temperatures (Secor and Gunderson 1998). The effects of temperature on green sturgeon, Acipenser medirostris, growth and bioenergetics have also been examined (Mayfield and Cech 2004; Allen et al. 2006). Mayfield and Cech (2004) also used thermal preference data to estimate lethal temperatures for green sturgeon. Studies of shortnose sturgeon have examined effects of temperature on spawning activity, egg development, and seasonal movements (Hall et al. 1991; Collins and Smith 1993; Kynard 1997; Collins et al. 2002; Hardy and Litvak 2004; DeVries 2006).

Despite the threats that summer water temperatures may pose to southern populations of shortnose sturgeon, few studies have examined temperature thresholds for shortnose sturgeon. Ziegeweid (2006) conducted laboratory experiments to determine LC50 temperatures for YOY shortnose sturgeon that ranged from 68 to 128 mm total length (TL) and were acclimated to a temperature of 19°C. Temperature LC50 values ranged from 28.2°C (±1.8°C) to 30.7°C (±0.8°C) (Ziegeweid 2006). Because acclimation temperature has been shown to affect thermal tolerance (Brett 1952; Becker and Genoway 1979; Jobling 1981; Kilgour and McCauley 1986; Young and Cech 1996), the results of Ziegeweid (2006) need to be re-evaluated with fish acclimated to a range of temperatures.

In this study, thermal maximum experiments were conducted with artificially propagated fish to examine how temperature tolerances of YOY shortnose sturgeon change with changes in acclimation temperature (Becker and Genoway 1979; Young and Cech 1996). Furthermore, thermal maximum data were used to estimate upper limits of safe temperature (Bridges 1971; Young and Cech 1996), final thermal preferences, and optimum growth temperatures for YOY shortnose sturgeon (Jobling 1981; Young and Cech 1996). Upon completion of experiments, total length, weight, and hematocrit levels of the fish at each acclimation temperature were measured to evaluate physiological condition of the fish in each treatment.

Materials and methods

Fish culture

Hatchery-produced young-of-the-year (YOY) shortnose sturgeon of the 2006 year class were raised at the Warm Springs National Fish Hatchery in Warm Springs, Georgia. Fish were raised in two 120-l cylindrical, flow-through tanks. The tanks were supplied with un-chlorinated spring water via polyvinyl chloride (PVC) pipes elevated about 15 cm above the tank. From this height, the falling water sufficiently mixed and oxygenated the water inside the tank. Suitable hardness, alkalinity, and pH were maintained in the source water with a chemical injection system (Hickson et al. 2001). Fish were fed a 3.0% body weight day−1 ration of Rangen® Soft-Moist commercial pellet feed (44% protein, 18% fat, <5% fiber, <8% ash, ~23% moisture). Tanks were siphoned at least once daily to remove accumulated waste and scrubbed at least once per week to remove algal growth on tank surfaces. A 14 h light:10 h dark photoperiod was maintained throughout the study.

Water in the two holding tanks was kept at two different experimental temperatures. Water in the first tank was maintained at 19.5°C, with an approximate flow rate of 85 ml s−1. The second tank was equipped with a 250-W Won® titanium aquarium heater and kept at a flow rate of 10 ml s−1 to maintain an average temperature of 24.1°C. This tank was further equipped with an airstone (connected to a central blower system) to help maintain suitable oxygen levels (Jenkins et al. 1993; Campbell and Goodman 2004) at lower flow and a higher temperature. The difference in flow rates was necessary to maintain the acclimation temperatures in the two treatments. However, because water was released straight down into the tanks, the incoming water did not create a current for the fish to swim against and thus should not have affected the results of the experiments. Dissolved oxygen and temperature were checked daily with a YSI-85® multi-parameter meter; other water quality variables such as hardness, alkalinity, and pH were checked weekly with colorimetric test kits (Aquarium Pharmaceuticals, Inc.).

Experimental setup

Thermal maximum experiments were conducted in two 38-l glass aquaria. Each aquarium was wrapped in a layer of clear plastic and a layer of 1.3-cm Styrofoam to minimize heat loss during experimental trials. Each aquarium was equipped with a 250-W Won® titanium aquarium heater, and water in each aquarium was circulated and oxygenated with an AquaClear Mini® biofiltration unit. Aquaria were elevated on 10-cm cinder blocks placed inside a 550-l cylindrical flow-through tank, which was used as a water bath to maintain desired water temperatures. The building housing the tanks was not temperature-controlled; therefore, the water bath was necessary to prevent the experimental aquaria from equilibrating with the ambient air. One of the 38-l aquaria was maintained at a temperature of 19.9°C, the approximate temperature of the source water. The other aquarium was maintained at 24.7°C with the aquarium heater. Both aquarium tanks were within 0.6°C of the acclimation holding tanks.

Experimental protocol

Juvenile shortnose sturgeon were held at their acclimation temperature for at least 7 days and starved for 24 h prior to use in thermal maximum experiments. In preparation for each trial, one fish from each holding tank was randomly selected and placed in the corresponding experimental aquarium. Fish were given 15 h to acclimate to the new tank conditions and recover from handling stress.

At the start of each trial, the water bath was drained, and both heaters were turned on. Water in the tanks was heated at a consistent rate of 0.1°C min−1 (Becker and Genoway 1979; Young and Cech 1996). Temperatures in each tank were continuously monitored with one of two YSI® meters, models 85 and 58. There was a dedicated meter for each tank so that experiments could be conducted for each acclimation temperature simultaneously. Prior to beginning thermal maximum testing, the two meters were compared against each other and consistently read within 0.1°C of each other. The starting temperature of each tank was recorded, and temperatures were recorded every 10 min and at designated endpoints (loss of equilibrium and death).

Because shortnose sturgeons are demersal, both loss of equilibrium (LOE) and death were designated as endpoints. Fish were considered to have lost equilibrium when they were unable to right themselves within 10 s of first losing equilibrium. Fish were declared dead when opercular movements ceased and the fish did not respond to tactile stimulus. In accordance with Becker and Genoway (1979), LOE endpoints are referred to as critical thermal maxima (CTmax), and death endpoints are lethal thermal maxima (LTmax).

Upon completion of each trial, mortalities were removed from the experimental tanks, total lengths and weights were recorded, and tank water quality measurements were recorded. Dissolved oxygen concentrations were measured with a YSI-85® multi-parameter meter, and hardness, alkalinity, pH, ammonia, nitrite, and nitrate levels were determined by using colorimetric test kits (Aquarium Pharmaceuticals, Inc.). Finally, the exchange rate of the biofiltration unit was estimated by using a 1.0-l beaker to collect and measure the water leaving the unit for a period of 10 s.

Once the trials for fish at both acclimation treatments were completed, the equipment was turned off, the aquaria were cleaned with a mild disinfectant, rinsed, and refilled, and the equipment was reset in preparation for the next trial. Trials were repeated throughout the summer as the fish grew and developed. A total of 32 fish, ranging in age from 64 to 140 dph, were tested at each acclimation temperature. Weights ranged from 0.6 to 35.0 g, and total lengths ranged from 54 to 215 mm.

Blood analysis

At the conclusion of all thermal maximum experiments, remaining fish in the holding tanks were removed, anesthetized in 75 mg l−1 MS-222 (Hickson et al. 2001), and measured for total length (mm) and wet weight (g). Blood was collected in heparinized 75-μl microcapillary tubes through caudal puncture. The tubes were sealed and centrifuged at 11 500 rpm for 10 min in an International Equipment Company (IEC) microcapillary centrifuge, model MB®. Hematocrit values were determined with a microhematocrit reader.

Statistical analysis

The normality of the distributions and the equality of the variances for the LOE and mortality data were examined using Shapiro–Wilkes and F max tests, respectively (Sokal and Rohlf 1981). Results indicated that the data were normally distributed and that variances were not significantly different between treatments (α = 0.05). In addition, sample sizes were relatively large (n ≥ 29). Therefore, parametric statistical tests were used to analyze the data. Lethal thermal maximum data were compared using ANCOVA, with acclimation temperature as the categorical predictor variable and the log10 (fish weight) as the continuous predictor variable. Weight data were linearized with log transformations because the rapid weight gain of YOY sturgeon skewed the distribution of the data. Upper limits of safe temperature (ULST) were determined by subtracting a safety factor of 5°C from the lethal and critical thermal maxima data (Bridges 1971; Young and Cech 1996). Final thermal preferences (FTP) and thermal growth optima (TGO) were estimated from CTmax data with regressions modified from Jobling (1981):

$${\text{FTP}} = \left( {{\text{CT}}_{\max } - 16.43} \right)\left( {0.66^{ - 1} } \right)$$
(1)
$${\text{TGO}} = \left( {{\text{CT}}_{\max} - 13.81} \right)\left( {0.76^{ - 1} } \right)$$
(2)

Critical thermal maxima, holding tank water quality, experimental water quality, and post-experiment length, weight, and hematocrit data were compared between the two acclimation groups with two-sample t-tests (α = 0.05).

Results

Water quality

Water quality variables were maintained at adequate levels in the holding tanks. Temperatures were significantly different between holding tanks (p < 0.0001), with mean temperatures of 19.5°C (SE = 0.1, n = 91) and 24.1°C (SE = 0.1, n = 91). Dissolved oxygen concentrations were significantly higher for the low acclimation temperatures (p < 0.0001), but the concentrations were adequate for both the low (8.3 mg l−1, SE = 0.1, n = 79) and high (7.4 mg l−1, SE = 0.1, n = 79) acclimation temperature treatments (Jenkins et al. 1993; Campbell and Goodman 2004). Mean hardness, alkalinity, and pH values for the holding tanks were 102.9 ppm (SE = 4.2, n = 60), 64.7 ppm (SE = 2.4, n = 52), and 6.9 (SE = 0.0, n = 60), respectively.

For thermal maxima trials, water quality variables were similar between experimental tanks for each acclimation temperature (Table 1). Furthermore, water quality variables in experimental tanks were maintained at levels that would not cause additional stress to the experimental fish (Table 1) (Piper et al. 1989). Ammonia, nitrite, and nitrate were rarely detected; if detected, concentrations did not exceed 0.25, 0.25, and 5.0 ppm, respectively.

Table 1 Water quality variables for experimental tanks used to conduct thermal maximum experiments with YOY shortnose sturgeon
Full size table

Fish condition

Total length (p = 0.6847) and weight (p = 0.5703) of experimental fish were not significantly different between acclimation temperature treatments. The mean total length and weight of fish used in thermal maxima trials were 135 mm (SE = 5, n = 64) and 10.7 g (SE = 1.2, n = 64), respectively. Furthermore, total lengths (p = 0.8809) and weights (p = 0.8638) of remaining holding tank fish were not significantly different between acclimation temperatures, with averages of 182 mm (SE = 2, n = 46) and 21.1 g (SE = 0.8, n = 46). Finally, hematocrit levels of remaining holding tank fish were not significantly different between acclimation temperatures (p = 0.9896), which demonstrated that oxygen requirements and general physiological condition were not affected by acclimation temperature.

Thermal maxima

The mean CTmax for fish acclimated to 24.1°C was significantly (p < 0.0001) higher (35.1°C, SE = 0.2, n = 32) than the mean CTmax for the fish acclimated to 19.5°C (33.7°C, SE = 0.3, n = 29) (Table 2). For LTmax data, the effects of acclimation temperature and the log10(fish weight), as well as the interaction between acclimation temperature and log10(fish weight), were significant (R2 = 0.87, p < 0.0001), and the following equations describe the relationships of the predictor variables:

$${\text{Low Temperature}}\left( {19.5^\circ {\text{C}}} \right){\text{LT}}_{\max } = 35.19 + 0.30\left[ {\log _{10} \left( {{\text{fish weight}}} \right)} \right]$$
(3)
$${\text{High Temperature}}\left( {24.1^\circ C} \right){\text{LT}}_{\max } = 35.82 + 0.49\left[ {\log _{10} \left( {{\text{fish weight}}} \right)} \right]$$
(4)

The distribution of LTmax data for fish at each acclimation temperature is shown in Fig. 1. Mean LTmax values were 36.1°C (SE = 0.1, n = 32) and 34.8°C (SE = 0.1, n = 32) for YOY shortnose sturgeon acclimated to temperatures of 24.1 and 19.5°C, respectively (Table 2). The sample size for the CTmax estimates is different for fish acclimated to 19.5°C because three of the fish in the low temperature trials died without first losing equilibrium. Estimated final thermal preferences and thermal growth optima were 26°C for the low acclimation temperature fish and 28°C for the high temperature fish (Table 2). Estimated upper limits of safe temperature ranged from 28.7°C to 31.1°C across critical and lethal thermal maximum data (Table 2).

Fig. 1
figure 1

Lethal thermal maxima distribution for YOY shortnose sturgeon raised at two different acclimation temperatures

Full size image
Table 2 Lethal thermal maxima (LTmax), upper limits of safe temperature (ULST), critical thermal maxima (CTmax), final thermal preferences (FTP), and thermal growth optima (TGO) for YOY shortnose sturgeon
Full size table

Discussion

Thermal tolerances of YOY shortnose sturgeon were significantly affected by acclimation temperature and body weight. A 4.6°C increase in acclimation temperature resulted in a 1.4°C increase in mean CTmax. A 4.6°C increase in acclimation temperature also resulted in a 1.3°C increase in mean LTmax, which indicated that death occurred at a relatively constant interval following LOE. In addition to acclimation temperature, LTmax values also were significantly affected by log10(fish weight) and by the interaction of log10(fish weight) and acclimation temperature. The LTmax increased with increasing weight, although the magnitude of this increase was greater for fish acclimated to the higher temperature (24.1°C).

Visual observations suggest that fish exhibited similar behaviors with increasing temperature regardless of acclimation temperature. As temperatures increased, fish activity appeared to increase. About 5–6°C prior to the lethal endpoint, fish began frantically swimming around the tank, presumably looking for an escape route. As fish began to lose equilibrium, their activity level decreased dramatically, and at about 0.3°C before the lethal endpoint, most fish were completely incapacitated.

The results of this study were similar to results of LC50 experiments for fish acclimated to a temperature of 18.8°C (Ziegeweid 2006). In the three LC50 experiments (Ziegeweid 2006), the average lengths of the fish were 68 mm TL, 106 mm TL, and 128 mm TL. In the LC50 experiments, the highest temperatures at which mortalities did not occur increased with size and were 26.5, 28.5, and 30.0°C, respectively. The lowest temperatures at which mortalities did occur were 30.1, 30.0, and 31.5°C, respectively. The averages of these described values create an interval of (28.3°C, 30.5°C) that can approximate an upper limit of safe temperature (ULST) for YOY shortnose sturgeon within the tested size range. Although the estimated interval is a function of the test temperatures, the narrow spacing of the test temperatures (1.5–3.5°C) and the abrupt change in mortality between test temperatures facilitate comparison to the ULST data from this study. Upper limits of safe temperature based on LOE data in this study were 28.7 and 30.1°C for fish acclimated to 19.5 and 24.1°C, respectively. Upper limits of safe temperature based on mortality data in this study were 29.8 and 31.1°C for fish acclimated to 19.5 and 24.1°C, respectively. Upper limits of safe temperature between the present study and Ziegeweid (2006) were very similar, despite differences in acclimation temperatures, measured endpoints, and methodologies.

The present study further illustrates the importance of acclimation temperature in evaluating thermal tolerance, which was not evaluated in the previous study of shortnose sturgeon thermal tolerance (Ziegeweid 2006). Thermal tolerance of YOY shortnose sturgeon increased with increasing acclimation temperature, which is similar to results of the euryhaline splittail, Pogonichthys macrolepidotus (Young and Cech 1996). However, the increase in thermal tolerance was not as pronounced in shortnose sturgeon. In this study, a 4.6°C increase in acclimation temperature only resulted in a 1.3°C increase in lethal temperature, whereas a comparable increase in splittail acclimation temperature (5°C) resulted in an 8°C increase in LOE temperature (Young and Cech 1996).

Although the thermal maximum method can be used to produce consistent, cost-effective results, the method has limitations. Final tolerances are influenced by acclimation temperature, thermal history, and heating rate (Becker and Genoway 1979; Jobling 1981; Kilgour and McCauley 1986), and the results of this study demonstrate that acclimation temperature significantly affects thermal tolerance. A heating rate that is too fast will result in thermal shock, whereas a heating rate that is too slow may result in partial acclimation of the fish, biasing the tolerance limit upward. The heating rate should be just fast enough to allow deep-body temperatures to parallel test temperatures without a significant time lag (Becker and Genoway 1979). The time lag for deep-body temperatures depends on the size of the test fish, but a heating rate of <1°C min−1 typically will not result in a significant time lag (Becker and Genoway 1979).

Concerns about heating rates in thermal maximum experiments can be addressed by conducting studies with a variety of heating rates. Studies of juvenile coho salmon, Oncorhynchus kisutch and pumpkinseed sunfish, Lepomis gibbosus, heated at a variety of rates ranging from 0.017 to 1.0°C min−1, suggest an optimum heating rate of 0.3°C min−1 (Becker and Genoway 1979); however, a heating rate of 0.1°C min−1 did not significantly bias the thermal maxima upward (Becker and Genoway 1979). Because coho salmon and pumpkinseed sunfish differ substantially in life history and body shape, a heating rate ranging from 0.1 to 0.3°C min−1 should be adequate for most small fish. Therefore, we assumed that the heating rate of 0.1°C min−1 used in this study did not affect the thermal tolerances of YOY shortnose sturgeon.

Published equations were used to estimate final thermal preferences (FTP) and thermal growth optima (TGO) for YOY shortnose sturgeon (Jobling 1981; Young and Cech 1996). The FTP estimates were similar to TGO estimates of fish acclimated to the same temperature, with values of approximately 26 and 28°C for fish acclimated to 19.5 and 24.1°C, respectively. The FTP and TGO estimates were close to the estimated upper limits of safe temperatures of 28.7 and 30.1°C for fish acclimated to 19.5 and 24.1°C, respectively. Similar results were observed for YOY splittail (Young and Cech 1996). Warmer temperatures may increase YOY metabolism and conversion of prey into somatic growth, as long as the cost of maintenance metabolism does not significantly increase (Brown 1957).

Temperature optima of many fishes are highest during early life history stages and decrease with ontogeny (Brett and Groves 1979; Kitchell 1979; Lankford and Targett 1994). For example, CTmax, ULST, FTP, and TGO decreased with increasing age for splittails acclimated to 17°C (Young and Cech 1996). In addition, juvenile green sturgeon grew better at 24°C compared to 19°C (Allen et al. 2006), even though a previous study had estimated an upper temperature limit of 27°C and reported significant transport-related mortality for fish acclimated to 25°C (Mayfield and Cech 2004). Similar trends in thermal tolerance are possible for shortnose sturgeon, although thermal maximum experiments with older juveniles are needed to confirm this relationship.

The thermal tolerance results of this study have important implications for wild populations of shortnose sturgeon, especially in southern rivers where summer water temperatures may limit habitat availability. Summer water temperatures in Georgia’s coastal rivers frequently exceed 31°C (Dyar and Alhadeff 1997; DeVries 2006), which would be potentially lethal according to the results presented in this study. Furthermore, at high temperatures, food availability may not be able to keep up with the metabolic needs of YOY shortnose sturgeon. Therefore, availability of thermal refugia in southeastern rivers may be critical in assessing shortnose sturgeon habitat in southeastern rivers. As such, identification of thermal refugia may be an important study component for management plans that seek to improve shortnose sturgeon recruitment.

Conclusion

The results of this study further demonstrate the importance of acclimation temperature in assessing thermal tolerances of fishes. Although the oxygen requirements and growth of YOY shortnose sturgeon were not affected by differences in acclimation temperature, thermal tolerances increased with increasing acclimation temperature. Estimated upper limits of safe temperature (ULST) ranged from 28.7 to 31.1°C and varied with acclimation temperature and measured endpoint. Final thermal preference and thermal growth optima were nearly identical for fish at each acclimation temperature and ranged from 26.2 to 28.3°C. Thermal maxima ranged from 33.7 to 36.1°C and varied with acclimation temperature and designated endpoint. The thermal tolerances identified in this study suggest a range of physiological limits on suitable habitats of YOY shortnose sturgeon in coastal rivers of the southeastern United States. This information should help to structure future investigations of YOY shortnose sturgeon that evaluate the hypotheses advanced in this discussion. Field collections of YOY shortnose sturgeons should be used to facilitate identification of critical habitats for YOY shortnose sturgeon in southeastern rivers. Furthermore, results of this study may be used to help determine appropriate mitigation measures for future hydrologic modifications that alter thermal regimes of southeastern rivers.