Turbulence effect on survival and feeding of Pacific bluefin tuna Thunnus orientalis larvae, on the basis of a rearing experiment


*Tel: 81-3-5351-6505.
Fax: 81-3-5351-6508. Email: y-kato@ori.u-tokyo.ac.jp


ABSTRACT:  Physical conditions such as oceanic turbulence related to food availability are considered to be important factors affecting fish larval survival. Rearing experiments were conducted to elucidate the effects of turbulence on the survival and feeding rates during the initial feeding period of Pacific bluefin tuna Thunnus orientalis. Six levels of turbulence intensity were provided by changing flow rates from pipes set on the bottom of rearing tanks. The result showed a dome-shaped relationship between turbulence level and survival rate, in which the feeding rate appeared higher at a logged turbulence energy dissipation rate of −6.32, and decreased at both higher and lower turbulence levels. Compared with the turbulence intensity in the ocean, the optimal turbulence level for Pacific bluefin tuna larvae corresponded to the turbulence caused by sea surface winds with speeds of 4–12.5 m/s. The estimated optimal turbulence intensity for Pacific bluefin tuna larvae is comparable to that for yellowfin tuna Thunnus albacares.


Pacific bluefin tuna Thunnus orientalis is a commercially very important fish species, particularly in Japan. Its consumption is increasing. However, stocks seem to be uncertain. Under this situation, sustainable fishing of this species is critical. Therefore, immediate settlement of this issue is a serious problem for not only Japanese fisheries but for all countries adjacent to the Pacific Ocean. Analysis of ovaries1 showed Pacific bluefin tuna spawns in a considerably limited area between Okinawa Prefecture and Taiwan during a limited season from late April to mid-June compared with other tuna genera. This specific spawning behavior increases the possibility that environmental factors in spawning areas affect survival and growth conditions in the initial life stage. Analyses of historical Pacific bluefin tuna fisheries catch data suggest its cohort strength would be determined in the early life stage until seven to eight months.2 Since only larger and faster-growing larvae were able to survive to the postflexion phase,3 consumption of an appropriate quantity of good-quality food is a critical factor for survival of initial feeding larvae, which do not have large energy stores.4

As one of the environmental factors controlling the initial survival rate, oceanic turbulence increases the encounter rate between planktonic predators and their prey. This effect has been well discussed in relation to the possibility of ingestion. Rothschild and Osborn5 estimated the components of predator and prey contact rate that are due to small-scale turbulence. This idea is originally described as an ‘optimal environmental window’ to indicate a relationship between fish recruitment and upwelling.6 The influence of wind- and tide-induced turbulence on encounter rates between larval fish and their zooplankton prey was quantified in numerical simulation models.7–10 These model results were confirmed by laboratory experiments using rearing tanks.11,12 In field studies, Sundby and Fossum13 indicated that small-scale turbulence and contact rates would be important regulatory mechanisms in the formation of year-class strength. Vertical tows with a multiple opening and closing net with an enviromental sensing system (MOCNESS) on the southern flank of Georges Bank were used to find the relationship between larval feeding success (of cod and haddock) and turbulent dissipation in a stratified water column.14 Lough and Mountain14 concluded that maximum feeding ratios occur at low and intermediate levels of turbulence where average prey density is greater than 10–20 prey/L. Utne-Palm and Stiansen15 and Utne-Palm16 investigated the effects of ontogeny, light, turbidity and turbulence on the attack rate and swimming activity of herring larvae and observed a dome-shaped effect of turbulence on attack rate. However, since most of these studies focused on larger larvae with an active swimming behavior, it is difficult to compare with the results of encounter rate between planktonic predators with little swimming ability and their prey.

To confirm the oceanic turbulence effects on the fish feeding rate, the initial stage fish larvae with little swimming ability should be used. Since the initial feeding behavior of these small fish larvae is related to oceanic turbulence, which affects their subsequent survival and growth, it is necessary to clarify the feeding behavior for development of culture techniques. Unsuccessful ingestion of available diets after yolk absorption probably results in abnormal behavior and morphological development, degeneration of the alimentary tract and trunk musculature, and reduction in food use efficiency and feeding activity.17,18 Kimura et al.19 found an optimal turbulence intensity that enhances larval survival of yellowfin tuna Thunnus albacares in the initial life stage, and suggested that the surface mixed layer of the equatorial ocean is used for spawning by this species is a suitable environment for larval survival. The study19 on the initial survival of fishery-important species indicates that oceanic turbulence should be considered as a physical factor controlling fish survival. Thus, in the present study, we focused on the effects of physical factors on Pacific bluefin tuna larvae using a newly developed experimental method and compared results obtained with results obtained from yellowfin tuna.19


Five-day rearing experiments from day 0 to day 4 were conducted at the Amami Station, National Center for Stock Enhancement, Fisheries Research Agency, Japan in July 2004. At the beginning of the experiments, 10 000 hatched Pacific bluefin tuna larvae were put into each 500-L tank. Three tanks were prepared for each turbulence level experiment. The tanks were set under natural light conditions. During the day, light was 1500–2500 lx. Water temperatures and dissolved oxygen were measured every morning and kept at 27–28°C and 6–7 mg/L. L-type rotifers Brachionus calyciflorus were put into the tanks as a larval diet from day 2 and maintained at a density of 5 inds/mL in each tank. Prey densities were measured twice a day at five points in each tank (near air bubbles and near the wall and center of the tank at two levels). Twenty larvae from each tank were collected and anesthetized with meta-aminobenzoic acid ethylester methanesulfonate (MS222, Sankyo, Tokyo, Japan) at 13:00 hours on day 4. The number of rotifers in the digestive tracts of larvae was counted under a microscope. The larvae were pressed down under a cover glass on a glass slide as soon as possible to remove influence of sampling lag. The number of surviving larvae was counted at the end of the experiments in each tank to estimate survival rate during the five-day rearing experiments.

Six levels of turbulence intensity were set by changing the flow rate of water pumped through pipes set on the bottom of the tanks (Fig. 1a). In addition, tanks using only air bubbles at 100 mL/min were also prepared as a control. The turbulence levels were set at 7.5, 15, 18, 30 and 48 L of water discharge per minute.

Figure 1.

(a) Schematic view of a rearing tank and turbulence generating apparatus. (b) Measurement points (●) for current velocity in rearing experiments and layers of measurement.

To compare these rates with those in the ocean, turbulence energy dissipation rates ε (m2/s3) in the rearing tanks were calculated by equation 1:12


where b is the intercept on the linear regression assuming a slope of −5/3, B is a constant, and the urms is the root mean square of the turbulent velocity. urms was measured by an acoustic Doppler velocity meter (Nortek, Vangkroken, Norway) at 12 points (six points for each layer) in the tank (Fig. 1b). The constant for B is chosen as B = 1 following by Stiansen and Sundby.12

Figure 2 shows the relationship between flow rates and logged turbulence dissipation rates. The dissipation rate ranged from 10−7.4 to 10−5.8 m2/s3, and the relationship was characterized by a linear regression (Pearson's correlation, r = 0.923, P < 0.05), while the turbulence dissipation rate at 7.5 L/min was slightly lower than the regression line. It is very reasonable that the logged turbulence dissipation rates have a linear relationship with flow rate because the relationship is similar to that between velocity and logged kinetic energy. Accordingly, six turbulence levels (named Control, Semi-Low, Low, Mid, Semi-High and High) were characterized for the turbulence dissipation rates (Fig. 2).

Figure 2.

Relationship between logged turbulence energy dissipation rate and flow rate. Line indicates linear regression.


Survival rates, defined as the number of larvae surviving in four days post hatch, are shown in Figure 3. The highest survival rate of 76.3 ± 5.5% (mean ± standard error) was recognized at ‘Semi-High’. The lower four levels without ‘Semi-Low’ showed considerably smaller survival rates around 25%, and most larvae could not survive at ‘High’.

Figure 3.

Relationship between turbulence levels and survival rate. Bars show standard error. Significant differences (Turkey's test, P < 0.05) between turbulence levels are indicated by different letters.

Figure 4 shows the number of rotifers in digestive tracts and the percentage of larval gut emptiness at each turbulence level. The percentage of larval gut emptiness refers to the ratio of number of larvae that can not ingest rotifers at all against the total number of larvae. The number of rotifers in digestive tracts increased gradually from ‘Control’ to ‘Semi-High’ with a peak at ‘Semi-High’. The number of rotifers in digestive tracts at ‘Semi-High’ was significantly higher than the others. The percentages of larval gut emptiness were the highest at ‘High’. Since this result means that few larvae in the turbulence level feed any rotifers, it is natural that both the number of rotifers in digestive tracts and the survival rates of larvae at that level were very low. These results suggest that the encounter rate of the larvae's diet becomes an important factor for feeding and that most larvae at ‘Semi-High’, where the logged turbulence energy dissipation rate is −6.32, can feed diets appropriately with a little feeding effort.

Figure 4.

Number of rotifers in digestive tracts (dashed line, ▴) and percentage of gut emptiness (solid line, ○) depending on turbulence levels. Bars, standard error. Significant differences (Turkey's test P < 0.05) between turbulence levels are indicated by different letters.


Pacific bluefin tuna larvae inhabit a layer shallower than 30 m depth in their spawning area south-west of Okinawa Prefecture where average water depth is greater than 1000 m.20,21 This indicates that tide-induced turbulence has little effect on bluefin larvae, while wind-induced turbulence can have a higher effect.

Figure 5 shows a vertical profile of dissipation rates dependent on wind-induced turbulence in the upper mixed layer for selected wind speeds. The estimated turbulence energy dissipation rates, ε, were calculated by equation 2:

Figure 5.

Relationship between depth and dissipation rate depending on wind speed (w). Gray shading indicates depth range of bluefin tuna larvae; black shading, optimum turbulence conditions and swimming depth range for bluefin tuna.


where w is the wind speed and z is the depth.22,23 According to Figure 5, the optimum turbulence intensity estimated in this study corresponds to the turbulence induced by wind at speeds of 4–12.5 m/s within the larval depth range. The results of our rearing experiments in this study suggest that wind mixing is necessary for larval survival. However, under strong turbulence conditions generated by extremely strong wind at speeds greater than 15 m/s corresponding to the ‘High’ turbulence intensity, most Pacific bluefin tuna larvae can not feed their diet and can not survive. In the spawning area, typhoons and large meteorological depressions often occur during the spawning season. These phenomena generate strong turbulent vertical mixing. This mixing is expected to affect the survival of Pacific bluefin tuna larvae including stock recruitment through the success or failure of appropriate ingestion of diet.

To confirm this hypothesis, the relationship between wind index and cohort strength of Pacific bluefin tuna2 was investigated (Fig. 6). The wind speed data of Miyako Island in Okinawa Prefecture, which is close to the spawning area, were obtained from the database of the Japan Metrological Agency (http://www.data.jma.go.jp/obd/stats/etrn/index.php). The wind index was defined as the sum of the number of days when the optimal wind speed of 4–12.5 m/s estimated in Figure 5 was observed during the spawning season (April–June).1,24 Although the Spearman rank correlation was calculated to be rs = 0.46 and the coefficient was low because of the small sample size, a positive correlation between wind index and cohort strength is shown in Figure 6. These results correspond with a previous study for the northern anchovy.25 Randall and Bradford25 found there was a significant linear relationship between larval mortality rate and the frequency of calm, low wind speed periods during the spawning season, suggesting that wind conditions control stock abundance through initial survival rate. However, many other physical factors such as water temperature and conditions of currents are related to wind conditions. Therefore, it is difficult to determine whether wind-induced turbulence directly affects larval survival. In particular, change in temperature caused by wind-induced vertical mixing should be considered in future studies.

Figure 6.

Relationship between normalized cohort strength (dashed line, ○) and wind index represented by the number of days with optimal wind conditions (solid line and ●). The number of days when the optimal wind speed was observed during this period was used to develop an index for wind conditions.

Let us compare the optimal turbulence intensity and wind conditions for Pacific bluefin larvae found in the present experiments with those for yellowfin tuna larvae in a previous study.19 Since Kimura et al.19 measured velocity in tanks once per second, it is impossible to directly compare their results with the turbulence dissipation rate calculated from spectral analyses carried out in the present study. Thus, velocity shear calculated by using equation 3 was used for this comparison (Fig. 7):

Figure 7.

Difference in survival rates for bluefin tuna (dashed line, ○) and yellowfin tuna (solid line, ●) dependent on vertical velocity shear.


where u′ is the standard deviation of velocity. Results showed that optimal turbulence intensity of yellowfin tuna was similar to that of Pacific bluefin tuna in the order of 10−7 m2/s3. Although this optimum range of turbulence intensity could be treated as showing the same characteristics for the two species, the peaks appeared at slightly different turbulence levels. This difference is probably attributed to air bubbles used in the yellowfin tuna experiment. In the previous study for yellowfin tuna, Kimura et al.19 used air bubbles for generating turbulence that might cause physical damage to the larvae. However, in the present study, this problem was avoided as turbulence was created only by water flow. To confirm this turbulence level as a common feature that affects tuna genera, further experiments using other tuna species are necessary.


We addressed the question of survival and feeding rate of Pacific bluefin tuna related to turbulence intensity. It was found that Pacific bluefin tuna larvae can not feed in strong turbulence in the initial feeding period when they have little swimming ability. There is an optimal turbulence that induced good feeding. This kind of study should be extended to other fish species, particularly coastal fishes.


Heartfelt thanks to Dr S Itoh and Dr H Kim, Ocean Research Institute, The University of Tokyo, for valuable comments. We also appreciate the cooperation of staff of the Amami Station, National Center for Stock Enhancement, Fisheries Research Agency and the Takuyo Corporation, who provided eggs for our experiments.