Pelagic fishes must obtain resources in prey-sparse habitats and may be considered energy speculators with maximization, gambling high energy costs (e.g. metabolism) for a high rate of return (prey capture). As such, they may have to carefully use their energy seascape to obtain the resources necessary for high growth rates. For diving animals, their energy seascape will also have a vertical component in addition to a horizontal one, which is rarely considered. Dolphinfish, Coryphaena hippurus, embody the maximization strategy as they have high metabolic rates and fast growth rates. We coupled biologging on wild individuals with lab-based respirometry to estimate dolphinfish swimming metabolic rates and vertical energy seascapes. Dolphinfish performed continuous yo-yo dives with deeper dives at night but higher activity during the day. Dive descents were ~27% less costly than the ascents. Fish modulated their behavior so that metabolic costs during the descent/ascent phases of deeper dives were less than those for shallow dives. While temperature is likely the primary limit of dive depth, the vertical energy seascape may be secondary and limit maximum dive depths. Studies of pelagic animal energy seascapes should consider the vertical component which will help determine their ability to access and utilize prey.

We coupled biologging on wild individuals with lab-based respirometry to estimate dolphinfish swimming metabolic rates and vertical energy seascapes. Wild-captured dolphinfish were used to directly relate fish activity to metabolic expenditure by placing ADL-tagged individuals into a large annular respirometer to measure oxygen consumption. Following the 24-h fasting period, the water level in the holding tank was decreased and fish were corralled using a soft, pliable vinyl barrier so that an individual fish could be guided into a water-filled sling. The captured fish was weighed, measured, and then fitted with a biologging tag package that included a multi-sensor ADL (model AGM, 6 cm in length, 2 cm wide, 1.5 cm high, 25 g in air, Technosmart, Rome, Italy) embedded in a buoyant float made of syntactic foam for tag recovery during field deployments. The ADL includes an accelerometer, gyroscope, and magnetometer each recording in all three axes (x, y, and z) at 25 Hz, as well as a depth and temperature sensor measuring at 1 Hz. Once the tag package was securely attached, the fish was moved into an 8851 L recirculating annular respirometer (3.05 m diameter, 1.22 m height) housed at the ARCC. Dissolved oxygen levels in the respirometer were recorded every 30-s using two oxygen sensors, one optical-type sensor (YSI Pro ODO, YSI, Yellow Springs, OH) attached through a port in the acrylic lid and one polarographic-type sensor (YSI Pro Plus Multimeter) in the recirculating loop. Routine metabolic rates (RMR; mgO2 kg-1 h-1) were calculated for each fish for a minimum of two-hour long segments from the measurement phases of the 48-h respirometry trials. The slope of the linear decline in oxygen consumption for each segment (R2 > 0.98) was multiplied by the volume of the respirometer minus the volume of the fish, divided by the mass of the fish and converted to units per hour.

Following the respirometry trial, acceleration data from the ADL were analyzed using Igor Pro v. 9.0.2.4 (WaveMetrics Inc., Lake Oswego, OR, USA) with the ‘Ethographer’ extension and R. We corrected for tag attachment angle by rotating the raw acceleration data until the x and y axis had a mean of zero. We removed the gravitational component of acceleration using a 3-s box smoothing window on the raw acceleration data. We used a continuous wavelet transformation with Ethographer to analyze cyclical patterns in the sway (lateral) axis of the accelerometers. This method classified the dominant peak as tailbeat cycles that were used to calculate tailbeat frequency (TBF) at 1-s intervals. TBF has previously been shown to be a good predictor for swimming speed and oxygen consumption of fishes. Mean TBF was calculated for each segment of oxygen consumption used to calculate RMR. TBF measurements less than 1.0 Hz were removed from analyses to avoid overestimating RMRs due to unnatural swimming behavior at low swim speeds typically yielding unnaturally high metabolic rates.

We used the respirometry-based relationship between TBF and RMR to estimate swimming metabolic rates (MRSW) of dolphinfish in the wild. Dolphinfish were tagged with ADLs off the coast of Baja California, Mexico and the Tsushima Strait, Japan due to an abundance of dolphinfish in these areas, and local resources to aid in tag deployment and recovery. We prepared the acceleration data from the field deployments the same way as the acceleration data from the lab calibrations described in the previous section. We also analyzed environmental data (depth and temperature) from the field deployments. To look at dive behavior, we categorized vertical swimming into three phases (ascent, descent, and level) using the fish’s vertical velocity (VV, m s-1), calculated as the difference between successive depth observations at 1 s intervals. Ascent and descent were defined as periods when the fish swam with VV less than or equal to -0.05 m s-1 or greater than or equal to 0.05 m s-1 for more than 10 s, respectively. Because dolphinfish have a high association with the surface [16, 17], we identified individual dive excursions as the period of time where a dolphinfish was ≥ 10 m in depth for longer than 10 seconds.