ORIGINAL ARTICLE

Energetics of swimming to shore in the puerulus stage of a spiny lobster: Can a postlarval lobster afford the cost of crossing the continental shelf?
John L. Wilkin 1 and Andrew G. Jeffs 2

Abstract Nonfeeding postlarval pueruli of spiny lobsters migrate tens of kilometers across the continental shelf to settle in coastal waters. A model that analyzes hydrodynamic forces during swimming in the puerulus of Jasus edwardsii is described. The model calculates the speed at which forward propulsion balances drag. Calculated speeds agree with observed puerulus behavior. The computed mechanical work is converted to metabolic energy consumption using an assumed efficiency. Values concur with reported estimates of the utilization of lipid energy reserves in pueruli. For biochemical energy reserves reported for pueruli collected 20 km off the east coast of New Zealand, the model suggests that this distance and durations of 5 days active swimming are the approximate limits to endurance. Sustained swimming exceeding 15 cm s21 will likely exhaust energy reserves before an animal can reach the coast, whereas swimming at less than 5 –7 cm s21 is inefficient because of the overhead of nonswimming, inactive metabolism. Successful onshore migration of this species is potentially limited by the animals’ energy reserves. Reduced energy reserves at the outset due to prior poor feeding, or delays encountered en route due to unfavorable currents, could lead to exceeding the stored reserves of the pueruli, and death. Potentially, relatively small shifts in coastal ocean climate conditions could generate marked changes in recruitment to important spiny lobster fisheries, as has recently been observed in many coastal populations of spiny lobsters.
Keywords: hydrodynamics, Jasus edwardsii, larval transport, New Zealand, swimming speed
1

Institute of Marine and Coastal Sciences, Rutgers, The State University of New Jersey, New Brunswick, New Jersey 08901, USA Leigh Marine Laboratory, University of Auckland, Warkworth 0941, New Zealand Correspondence to John L. Wilkin, jwilkin@rutgers.edu

2

Introduction [1] The postlarval or puerulus stage in the life cycle of the spiny lobster (Palinuridae) is a critical link between the pelagic larval development in offshore waters and benthic juvenile-adult phase in shallow inshore waters (Phillips et al. 2006a). Spiny lobsters are unique among ma-

rine organisms in that the nonfeeding pueruli migrate tens of kilometers across the continental shelf into shallow coastal waters, where they settle and become juveniles. The manner in which the pueruli move inshore is poorly understood, but there is evidence that this movement involves active swimming that utilizes

Limnology and Oceanography: Fluids and Environments † 1 (2011): 163–175 † DOI 10.1215/21573698-1504363 q 2011 by the Association for the Sciences of Limnology and Oceanography, Inc.

Downloaded at ASLO on January 1, 2012

164

Limnology and Oceanography: Fluids and Environments

1 (2011)

significant lipid reserves (Takahashi et al. 1994; Jeffs et al. 1999, 2005). These reserves have been built up over the long (9– 24 month) larval development period, depending on the species (Wells et al. 2001; Jeffs et al. 2004; Phillips et al. 2006a). In order to be able to establish successfully once settled in shallow coastal waters, the pueruli must have sufficient remaining biochemical energy reserves to molt into a benthic juvenile (Phleger et al. 2001; Jeffs et al. 2002). Consequently, it has been suggested that the exhaustion of puerulus energy reserves may play a role in determining the high interannual variability in recruitment that characterizes spiny lobster populations (Jeffs 2001; Jeffs et al. 2001a; Phillips et al. 2006b). It may also play a role in the recent dramatic decline in lobster fisheries throughout the world following more than 30 years of stability, as global perturbations in ocean and coastal climatic conditions may compromise puerulus energy accumulation through reductions in larval food availability or by altering water currents influencing onshore migration (Jeffs 2001, 2010; Caputi et al. 2010). [2] Demonstrating this method of migration by active swimming is problematic because pueruli are difficult to study in the field, due to their low abundance, wide dispersal, and highly cryptic behavior and appearance (Phillips et al. 2006a). Laboratory flume tank experiments intended to directly measure the energetic costs of swimming in the nektonic pueruli of New Zealand’s red spiny lobster, Jasus edwardsii, were previously thwarted by the clinging behavior of the pueruli when placed in the flume (Jeffs and Holland 2000; Jeffs 2001). However, estimates of the effort associated with the onshore migration of pueruli have been derived from measuring changes in the biochemical energy stores of pueruli caught at different distances offshore. This is possible because the puerulus stage does not feed and so relies entirely on stored energy in the form of lipid (Jeffs et al. 1999, 2001b; Phillips et al. 2006b). Comparison of the biochemical composition of pueruli arriving at the coast with those caught 20 km immediately offshore on the southeast coast of the North Island, New Zealand, showed marked differences in energy stores associated only with their lipid content (Jeffs et al. 1999, 2001b). The total consumption of lipid reserves in pueruli of J. edwardsii crossing this 20-km stretch of

the continental shelf was estimated at 3.1 ^ 1.0 [mean ^ SE] mg, which is equivalent to a total energy expenditure of 113 ^ 36 J (Jeffs et al. 1999). Over the 20 km this implies an average swimming effort of 5.6 ^ 1.8 J km21, but this figure would be somewhat lower if the pueruli do not follow a direct course. This energy consumption for migrating pueruli of J. edwardsii lies in the middle of the theoretical range (1– 10 J km21) estimated to be the net energetic cost of swimming for an organism the size of a puerulus (Morris et al. 1985) and calculated as the range for swimming in the Western Australian lobster, Panulirus cygnus (Lemmens 1994). This estimate of energetic cost was consistent with subsequent studies of pueruli of the same species over a greater range of distances offshore (up to 284 km offshore) (Chiswell and Booth 1999; Jeffs et al. 2001a). [3] Therefore, the objective of this study is to develop an independent estimate of the effort expended by pueruli during sustained swimming using a simple model of the hydrodynamics of their swimming motion. We hypothesize that this model can be used to predict distances and elapsed times for migrating pueruli of J. edwardsii, assuming a metabolic efficiency typical of that reported for similar crustacea. Furthermore, we test the veracity of the model by reconciling total energy requirements of migrating pueruli generated by the model with independent estimates of energy requirements in pueruli using biochemical measurements.
Model Description and Results Mechanics and Energetics of Swimming [4] Although it can be problematic to directly measure the metabolic power used in swimming by small crustaceans, estimates of swimming energetics can be made by considering simplified models of the hydrodynamic thrust and drag of the swimming appendages and body. Morris et al. (1985, 1990) successfully used this approach to estimate the propulsion efficiency, swimming speed, and jump distance of small cyclopoid and larger calanoid copepods. We used a similar approach to estimate the hydrodynamic power required to maintain a constant swimming velocity in the puerulus stage of the spiny lobster by calculating the propulsive and drag forces of the swimming appendages integrated over a complete swimming beat and balancing these with the

q 2011 by the Association for the Sciences of Limnology and Oceanography, Inc. / e-ISSN 2157-3698

Downloaded at ASLO on January 1, 2012

165

Larval lobster swimming energetics

Wilkin and Jeffs

steady drag force on the animal. The angular displacements and velocities of the pleopods (the swimming appendages beneath the tail of the puerulus) were obtained from video imagery, whereas the pleopod dimensions (which differ in effective hydrodynamic cross section between the power and recovery strokes) were measured from sample animals. For the present study, actively swimming pueruli of the spiny lobster J. edwardsii were gathered from puerulus collectors located in the Port of Gisborne, New Zealand (Booth 1994). [5] The copepods studied by Morris et al. (1985, 1990) accelerate significantly on each power stroke and decelerate on the recovery stroke, a factor they incorporated into their swimming model. In contrast, the lobster pueruli considered here are observed to maintain a constant velocity during sustained forward swimming (Jeffs and Holland 2000). This is consistent with their substantially greater mass (typically 0.3 g), and hence inertia, compared with small crustaceans such as copepods. The assumption of a steady swimming velocity is not valid for the rapid backward swimming behavior of pueruli exhibited during an escape reaction; however, our interests here are not in the swimming energetics during the evasion of predators but rather in the net force balance for sustained forward swimming during migration across the continental shelf to the coast. The assumption of a constant swimming velocity simplifies the swimming model compared with that of Morris et al. (1985, 1990) (for copepods), but the nonlinearity of the quadratic drag force on a puerulus-sized animal nevertheless necessitates an iterative solution for the swimming speed at which propulsive and drag forces are balanced.
Model of Swimming Energetics [6] Our swimming model comprises three basic elements: the propulsive forces generated by the regular beat of the swimming appendages, the drag on the appendages during the recovery stroke, and the drag on the body itself. When these forces are balanced, a steady swimming velocity is achieved. Inputs to the model are the pleopod dimensions, the observed angular movements of the pleopods, the beat rate of the swimming motion, and the dimensions of a typical puerulus body (which determines the drag).

Motion of the Swimming Appendages [7] The swimming behavior in the pueruli of J. edwardsii has been described previously by Jeffs and Holland (2000). Using video imaging, they examined an ensemble of some 30 pelagic pueruli held in a seawater tank within a few hours of capture and observed relatively uniform swimming velocities averaging around 16 cm s21. The four pairs of pleopods that propel the animal in steady swimming were seen to beat in synchronous metachronal waves at a rate of about 6.5 beats s21. [8] In our swimming model we describe the sweep of the puerulus pleopods in terms of the angle each pleopod makes to the direction of motion as a function of time, bi(t), for the ith pleopod, numbering from the head to the tail. The minimum and maximum angular displacements of the pleopods (Fig. 1) were estimated by analyzing the movements of six swimming pueruli using a high-speed video recorder (JVC Pro HD, Japan) with a macro lens through a glass tank (2.1 m long · 0.5 m wide · 0.75 m high) in a manner similar to that described by Jeffs and Holland (2000). Measurements for each of the four pleopods taken from six

VS

0

5 mm 1

10

Pleopod:

4

3

2

bmin bmax bmin bmax –9° –158° –17° –139° –17° –132° –16° –122° b(t )

Fig. 1 The pleopods of the puerulus of Jasus edwardsii that are actively used for swimming are numbered 1 through 4 from head to tail. The angular displacement at time t for each pleopod, b(t), is measured from the direction of travel, and the ranges of motion, bmin to bmax, for each pleopod are given. The puerulus swims forward at a steady speed VS.

q 2011 by the Association for the Sciences of Limnology and Oceanography, Inc. / e-ISSN 2157-3698

Downloaded at ASLO on January 1, 2012

166

Limnology and Oceanography: Fluids and Environments

1 (2011)

Forward Propulsion Force from Pleopod Motion [13] The propulsive force calculation is achieved by discretizing each pleopod along its length into a finite number of segments (we use NR ¼ 50), computing the force on each segment in accordance with its local velocity and 4 3 2 1 cross-section dimension, and 0 then summing over all segments. 1 This computation is repeated at a 2 finite number of time intervals 3 (we use NT ¼ 200) during the cycle of power and recovery 4 stroke, and the total propulsive 5 force is integrated over the time 6 –2 0 2 –2 0 2 –2 0 2 –2 0 2 of a complete beat. Pleopod width (mm) [14] The forces on a single pleopod segment during the Fig. 2 Idealized pleopod widths of the puerulus of Jasus edwardsii used in the swimming model. Light black line is power stroke at time t (Fig. 4) model profile on the power stroke; heavy red line is model profile on the recovery stroke.
Radial distance from joint (mm)

pueruli were averaged to give the parameters of the pleopod strokes in our swimming model. [9] The pleopods from two of these pueruli were removed and scale drawings made using a microscope and camera lucida. Measurements taken from the drawings determined the width of each appendage as a function of the distance along the pleopod, starting from the articulated joint with the abdomen (Fig. 2). In our video observations, the exopodites (the biramous terminal paddles on the pleopods; see Jeffs and Holland 2000, their figure 2) were seen to spread apart on the power stroke, while on the recovery stroke they were brought together. This action reduces the hydrodynamic cross section and decreases the drag on the recovery stroke. This difference in effective cross section on power and recovery strokes was determined from our measurements of the sample exopodites (Fig. 2). [10] The power and recovery strokes of the pleopod motion are separated by resting intervals at the change of stroke direction. The sum of stroke, recovery, and resting intervals comprises one beat (Fig. 3), and the duration of the respective components of the beat scale proportionately as the animal varies the beat frequency. [11] The video imagery of puerulus swimming action shows that on the power strokes the pleopods move separately and, we assume, develop their thrust independent of the motion of the adjacent pleopods. If there were partial interference in the flow around the pleopods due to wake turbulence, we would be presented with a virtually intractable hydrodynamic problem. We simply postulate here that any interference in

the hydrodynamics of the appendages would result in a decrease in overall swimming efficiency, and that this is unlikely in an animal for which efficient sustained swimming is critical to this particular life stage. The angular displacement versus time of each of the four pleopods was derived from the mean measures taken from video recordings of six actively swimming pueruli (Fig. 3). For convenience, a cosine function is adopted for the transition from starting to ending displacement, and back. Differentiating the angular displacement with respect to time gives the angular velocity, v, which enters in the swimming model. Each pleopod has essentially completed the stroke before the adjacent pleopod commences (Fig. 3), supporting the conjecture that there is minimal hydrodynamic interference of neighboring pleopods during the power stroke. The thrust developed by the four pairs of pleopods is therefore computed independently and summed to give the total propulsive force of the swimming action. [12] In contrast to the power strokes, the recovery stroke is characterized by a synchronized motion of the four pleopods, presumably to minimize drag, on the return. Accordingly, we compute the combined drag of all appendages on the recovery stroke as simply that for the pleopod experiencing the greatest drag during the course of the return movement.

q 2011 by the Association for the Sciences of Limnology and Oceanography, Inc. / e-ISSN 2157-3698

Downloaded at ASLO on January 1, 2012

167

Larval lobster swimming energetics

Wilkin and Jeffs

0 Angular displacement b (°)
4 3 21

–50
P4

P3

P2

P1

term and an inertia force associated with the effect of hydrodynamic added mass due to displacement of the water (Morison et al. 1950) during acceleration of the pleopods: dU ij 1 DDij ¼ rC D Aij U 2 þ rK A Lij ; ij 2 dt ð3Þ

–100 –150
Rest Power Rest Recovery

–200

0

0.2 0.4 0.6 0.8 t/TB = time during stroke/period

1

Fig. 3 Angular displacement b (see Fig. 1) of pleopods P1 through P4 versus time normalized by beat period TB for a puerulus of Jasus edwardsii. For pleopod 2 (heavy black line), the portions of the complete stroke that are the power, recovery, and rest periods are indicated.

where r is the density of seawater, CD is a drag coefficient, Aij is the projected cross-sectional area of the segment normal to the flow, KA is the added mass coefficient, and Lij is the volume of the segment. We again follow Morris et al. (1985) and use an empirical formula for the drag coefficient of a cylinder normal to the flow: C D ¼ 1 þ 10 Re22=3 ; ð4Þ

are defined in terms of the angular velocity of the ith pleopod, vi(t) ¼ dbi/dt, and the speed perpendicular to the leg associated with the sweep action, rijvi, where rij is the radial distance along the leg from the joint to segment j for pleopod i. To this we add the component of the steady swimming velocity (VS) that is perpendicular to the leg to obtain the actual speed, U, of segment j through the water: U ij ¼ r ij vi 2 V S sin bi : ð1Þ

The hydrodynamic drag force on this segment is denoted DDij, and its component in the direction of motion, DF ij ¼ DDij sin bi ; ð2Þ

is the thrust, DF, that the segment contributes to the animal’s propulsion. [15] Thus far we have followed rather closely the conceptual model of swimming dynamics proposed for copepods by Morris et al. (1985). However, unlike in their model, we need not consider unsteady movement of the body and the role that the cycle of acceleration and deceleration on each stroke plays in the hydrodynamic added mass of the animal. This is because our video observations show that puerulus swimming speeds are quite steady over the course of a beat, and the net acceleration is negligible (Jeffs and Holland 2000). [16] We compute the propulsive force on each swimming appendage as the sum of a quadratic drag

where Re ¼ Ud/n is the flow Reynolds number based on cylinder diameter, d, and n ¼ 1026 m2 s21 is the kinematic viscosity of water. Here, the cylinder diameter is dij, the width of the relevant pleopod segment (Fig. 2), recalling that this width changes between recovery and power strokes when the pair of overlapping exopodites on each pleopod separates. The area Aij in equation (3) is the product of dij and the small radial length of each segment, drij ¼ Li/NR, where Li is the full length of the ith pleopod. We continue with the assumption that the pleopod segment is approximately cylindrical to compute the volume of fluid displaced, Lij, and use KA ¼ 1 appropriate to a cylinder in an accelerating flow. It can be shown that the ratio of the inertia to drag forces scales with the aspect ratio (diameter/length) of the pleopod, so it is significantly less on the recovery

rij
b(t ) ∆Fij = ∆Dij sin b drij

∆Dij
Fig. 4 Schematic diagram of the forces on a pleopod segment drij during the power stroke. At angular displacement b, DFij is the component of thrust DDij in the direction of motion.

q 2011 by the Association for the Sciences of Limnology and Oceanography, Inc. / e-ISSN 2157-3698

Downloaded at ASLO on January 1, 2012

168

Limnology and Oceanography: Fluids and Environments

1 (2011)

stroke than on the power stroke. In our computations the added mass inertia is approximately 30% of the drag force on the power stroke. [17] The thrust generated by a single pleopod at time t is the summation over all segments along the length of the pleopod, F i ðtÞ ¼
NR X j¼1

F ij ¼

NR X j¼1

DDij sin bi ðtÞ;

ð5Þ

and is positive on the power stroke but negative, representing drag, on the recovery stroke. We calculate the total forward thrust achieved by the coordinated action of all pleopods by summing the efforts of the eight pleopods (four left and four right) and averaging over the duration of the power stroke, TP , that is, in the interval 0 , t , TP . As noted previously, video imaging shows that on the recovery stroke the pleopods tuck together and move synchronously, thereby shadowing each other and decreasing the extent of drag that would hamper forward motion. Therefore, on the recovery stroke the calculation must find the pleopod that is experiencing the greatest drag at any given time. On recovery the thrust Fi , 0, so the greatest magnitude of the drag is actually given by minimum over i of Fi(t) during TP , t , TB. The average thrust over a complete beat cycle is then ! ðTB 4 ð 2 X TP F i ðtÞdt þ min F i ðtÞdt ; ð6Þ F Thrust ¼ T B i¼1 0 TP i where the factor of 2 accounts for the combined effort of paired left and right pleopods. [18] In sustained, steady swimming this net propulsive force is opposed by drag on the body, which we parameterize by adopting a drag coefficient for the puerulus body using an empirical relation based on frontal area for a streamlined body of revolution (Blevins 1984): "  1=2 # L d C DjBody ¼ 3 £ 1:328 Re 21=2 þ d L ð7Þ d þ 0:33 ; L Re , 105 ;

of more than 100 pueruli collected from the southeast coast of the North Island of New Zealand, we obtained widths d ¼ 6.0 ^ 0.2 mm and lengths L ¼ 35.0 ^ 0.3 mm. Blevins (1984) notes that equation (7) indicates that the optimum aspect ratio for minimum drag is about L/d < 5, and the sample pueruli are close to this criterion. This adds credence to our choice of drag coefficient because we expect the puerulus stage to be adapted to minimize drag for efficient swimming in the migration to shore. [19] The drag force at swimming speed VS is then 1 p F Drag ¼ rC DjBody d 2 V 2 : S 2 4 ð8Þ

[20] An assumption embodied in equation (1) is that the segments of the pleopods developing the thrust lie outside the frictional boundary layer that forms around the puerulus body; otherwise, we would need to use a value smaller than VS when rij is less than the boundary layer thickness. We justify this by considering the Blasius solution for the thickness of the boundary pffiffiffiffiffiffiffiffiffiffiffi layer that grows on a flat plate: d ¼ 2 nx=V , where x is the distance from the leading edge (here, from the head of the puerulus). For VS ¼ 8 cm s21, d is of order 1 mm at the tail of the puerulus x ¼ L (35 mm), which indicates the swimming surfaces of the pleopods (Fig. 2) indeed lie outside the frictional boundary layer surrounding the body.
Steady Swimming Velocity [21] Whereas swimming velocity obviously affects the drag on the body, it also affects the total thrust generated by the beating pleopods via equation (1). The steady swimming speed achieved by a puerulus is therefore not an independent parameter of the swimming model, but rather the velocity the animal attains when drag matches the net forward thrust, F ¼ D. [22] It is not possible to combine equations (6) and (8) to derive an explicit algebraic expression for VS, but the condition FThrust ¼ FDrag is readily solved iteratively using a bisection algorithm to compute VS as a function of beat frequency (Fig. 5). The model gives steady swimming speeds of 13–16 cm s21 at beat frequencies of 6– 7 s21. This result is in close agreement with the measurement of actual steady swimming

where d is the maximum body width, L is the body length, and the Re is based on L. From measurements

q 2011 by the Association for the Sciences of Limnology and Oceanography, Inc. / e-ISSN 2157-3698

Downloaded at ASLO on January 1, 2012

169

Larval lobster swimming energetics

Wilkin and Jeffs

Steady swimming speed (m s–1)

0.3 0.25 0.2 0.15 0.1 0.05 0 0 5 10 Beat frequency (s–1)

Fig. 5 Steady swimming velocity for a puerulus of Jasus edwardsii as a function of beat frequency computed by solving for speed VS at which FThrust (equation (6)) ¼ FDrag (equation (8)). Red shading shows sensitivity to ^ 30% change in pleopod drag coefficient. Blue dashed line is sensitivity to ^30% variation in body drag coefficient.

speeds of 13.0 –30.7 cm s21 in freely swimming pueruli by Jeffs and Holland (2000) and provides further support of the validity of the swimming model.
Sensitivity to Parameters [23] A number of assumptions have been made in formulating the puerulus swimming model, not the least of which are that the pleopods have a simple rotational movement and that they may be treated, for hydrodynamic purposes, as piecewise cylinders with dimensions held fixed on the power and recovery strokes. Accepting for the moment that the model is conceptually sound, it is nonetheless prudent to examine how uncertainties in model parameters will carry through to the calculated net force balance, steady swimming velocity, and energetics. [24] The observed variance in the dimensions of bodies and appendages is remarkably low—at most a few percent (Jeffs et al. 1999). We also have confidence from the video imagery that the timing and angular displacements of pleopod motion are well measured. Uncertainty in the balance between propulsive and drag forces therefore arises principally from the poorly known drag coefficients in equations (6) and (8) and the simple treatment of added mass inertia in the pleopod thrust calculation.

[25] Whatever the actual source of uncertainty in the thrust calculation, we can explore its impact by allowing sizeable variation in CD in equation (3). The range in steady swimming speed is about ^8% when CD is varied by ^30% (shaded region in Fig. 5). This relative insensitivity arises because the total thrust is only partially dependent on CD (due to the independent added mass inertia term) and because the drag force that balances net thrust is proportional to the square of swimming speed. For example, a 17% increase in total thrust would be required to produce an 8% increase in swimming speed. Plotting the sensitivity to ^30% changes in CDj Body reveals that the effect on swimming speed is comparable to, but opposite, the pleopod drag (see dashed lines in Fig. 5). The swimming speed at fixed beat frequency is insensitive to KA because the inertia effect is a linear function of flow acceleration and is essentially symmetrical on power and recovery strokes. However, the animal must do work against this force in both stroke directions, so sensitivity to KA does become evident in net energy consumption—a point we return to below.
Energetics and Hydrodynamic Efficiency [26] To compute the effort expended in sustained swimming, we use the principle that the work done by the swimming appendages in the course of one complete pleopod beat is the product of force times distance swept by each pleopod, integrated over the duration of the beat. Recalling that the velocity of each pleopod segment is rijvi(t), the work done in a small time interval dt is jDDijrijvi(t)j, where we take the absolute value because the animal expends this effort regardless of which direction the pleopod is moving. The power provided by the average mechanical work (or energy) per time associated with swimming is then

PMech

4 2X ¼ T i¼1

ðTB X NR
0 j¼1

DDij r ij vi ðtÞ dt;

ð9Þ

where the summation over the eight pleopods (four left and four right) has been applied, and the division by beat period yields a result in units of power, that is, joules per second, or watts.

q 2011 by the Association for the Sciences of Limnology and Oceanography, Inc. / e-ISSN 2157-3698

Downloaded at ASLO on January 1, 2012

170

Limnology and Oceanography: Fluids and Environments

1 (2011)

[27] The fraction of the mechanical effort that is converted into useful power by overcoming drag on the body to the mechanical effort expended by the animal is the hydrodynamic efficiency:

mH ¼

F Drag V S : PMech

ð10Þ

Our swimming model indicates this ratio is relatively insensitive to swimming velocity, varying from 7% at VS ¼ 3 cm s21 to 6% at VS ¼ 20 cm s21. These values are smaller than the 30% estimated by Morris et al. (1985) from their swimming model of the small copepod Pleuromamma xiphias. Svetlichny and Hubareva (2005) measured respiration rates for food-deprived copepods tethered to a force sensor, and from their reported data we compute hydrodynamic efficiency for locomotion by mouth appendages of 7%, 13%, and 21%, respectively, for Paracalanus parvus, Pseudocalanus elongatus, and Calanus euxinus. Visser (2007) assumed a much smaller value for mH of 1% in a study of the energetics of zooplankton foraging behaviors. [28] To convert the rate of mechanical work to an estimate of metabolic effort, PMET, we assume the same muscle efficiency mM ¼ 0.25 proposed by Morris et al. (1985) for other crustacea and compute PMET ¼ m21 M PMECH.
Energy consumption rate (J h–1) from mechanical work/muscular efficiency 10

[29] A data set with which we may compare the swimming effort predicted by the model is field data gathered by Jeffs et al. (1999) on the utilization of biochemical energy stores by pueruli of this species migrating over a 20-km distance from offshore into shallow coastal waters. Pueruli of this species typically migrate over much larger distances from the point of metamorphosis from the prior larval stage (92.4 ^ 7.8 km; Jeffs et al. 2001a), and the energy consumption rates estimated by our model apply for the entire postmetamorphosis migration.
Discussion Energy Requirements of Pueruli [30] Estimates of rates of energy consumption from the swimming model were compared with independent biochemical measures of rates of energy consumption of puerulus of J. edwardsii over the last 20 km of migration. The model was used to address the question of whether a typical puerulus as examined in a previous study has sufficient energy reserves in stored lipid to afford the cost of swimming the 20 km across the shelf to settle at the coast (Jeffs et al. 1999). In our model the swimming effort can be moderated only by varying the frequency at which the pleopods beat, and naturally, the faster they beat, the greater the speed the animal can attain. However, hydrodynamic drag and decreased velocity of the pleopods relative to the surrounding water on the power stroke (equation (1)) mean that swimming ever faster incurs an increasing penalty in terms of energy efficiency. Consequently, metabolic effort due to swimming (i.e., PMET J h21) increases much more rapidly than does swimming speed (Fig. 6). Therefore, if metabolic cost were solely due to swimming and no other factors limit the duration of cross-shelf migration, the most efficient migration strategy would be to conserve energy by swimming very slowly. However, pueruli expend metabolic effort whether swimming or not, and this would exhaust their energy reserves before settlement if they made insufficient progress in their crossshelf migration. Lemmens (1994) measured the metabolic rate of recently settled pueruli (i.e., resting and not swimming) of the Western Australian spiny lobster, Panulirus cygnus, which enabled him to calculate their individual energy requirement at 0.54 J h21. The pueruli

8

6

4

2

0 0 0.05 0.1 0.15 0.2 Steady swimming speed (m s–1)
Fig. 6 Rate of metabolic energy consumption PMET as a function of steady swimming velocity in a puerulus of Jasus edwardsii (solid black line): sensitivity to ^ 30% variation in puerulus body drag coefficient (red shading), ^ 30% variation in added mass coefficient (blue shading), and ^30% variation in pleopod drag (green shading).

q 2011 by the Association for the Sciences of Limnology and Oceanography, Inc. / e-ISSN 2157-3698

Downloaded at ASLO on January 1, 2012

171

Larval lobster swimming energetics

Wilkin and Jeffs

of J. edwardsii are of a similar size as P. cygnus, so we will assume they have a similar basal metabolic energy requirement, which we denote PREST, that will add to the metabolic cost of active swimming. [31] The puerulus stage of spiny lobster species, including J. edwardsii, exhibits diurnal vertical migrating behavior, with active swimming only at night in coastal waters to avoid visual predators (Booth and Phillips 1994; Acosta and Butler 1999; Phillips et al. 2006a). If we denote by a the fraction of time a puerulus spends actively swimming, then in any 24 h the effective swimming speed is aV S and the average rate of energy consumption is PSWIM ¼ PREST þ am21 PMECH : M ð11Þ

25 Energy consumption rate (J km–1) Sustained swimming 20 50% resting

15

10

5

0 0 0.1 0.2 Speed during swimming (m s–1)
Fig. 7 Rates of energy consumption with (a ¼ 0.5, solid black line) and without (a ¼ 1, green dashed line) diurnal inactivity in swimming behavior, as a function of active swimming velocity. Red shading shows the sensitivity of PSWIM for a ¼ 0.5 to ^30% variation in puerulus body drag coefficient.

[32] At high swimming speeds the inclusion of diurnal inactivity has little effect on energy use per distance because active swimming dominates the effort, but at slower speeds diurnal inactivity affects the overall energy requirement because of the relative greater burden of the basal metabolism for each kilometer of progress (Fig. 7). When speeds fall below 5 cm s21 for sustained swimming (a ¼ 1), and below 7 cm s21 for nighttimeonly swimming (a ¼ 0.5), the energy use per distance traveled increases significantly. It could be argued that 5 – 7 cm s21 therefore represents an optimal swimming speed that offers the greatest endurance, if all other factors influencing the duration of the migration can be discounted. Certainly, the swimming model suggests that swimming speeds less than 5 cm s21 would be an inefficient use of energy. [33] The total energy consumed (PSWIM · duration) and distances traveled were calculated for a set of active swimming velocities of Vs ¼ 7.5, 15, and 20 cm s21, with a ¼ 0.5 (Fig. 8). Distance traveled is based on the average speed aVS. Swimming speeds of less than 7 cm s21 are not considered because we have demonstrated these would make poor use of the stored energy of a puerulus. These results indicate that 3 – 8 days are required to traverse the necessary distance (i.e., 20 km) at the range of nighttime-only swimming speeds considered (Fig. 8). The least energy required for a 20-km migration is some 135 J at 7.5 cm s21, and at that speed the journey takes 6 days. A more rapid transit

at around 15 cm s21 typical of the speed measured in free-swimming pueruli (Jeffs and Holland 2000) may be advantageous due to the elevated predation risk for pueruli traveling in coastal waters where visual-hunting pelagic predators are more abundant and to which they are known to be particularly prone (Gracia and Lozano 1980; Howard 1988; Acosta and Butler 1999). If warranted to avoid predation, this accelerated speed would incur a significantly higher energetic cost of some 190 J. Interestingly, an average speed of onshore migration in pueruli of J. edwardsii was estimated at 8– 10 cm s21 from catch distribution data of larvae and pueruli from a series of transects offshore of New Zealand (Chiswell and Booth 1999). This would be the equivalent of swimming at 16–20 cm s21 assuming a proportion of diurnal inactivity of a ¼ 0.5. [34] Our estimate using the swimming model that at least 135 J is required to execute a 20-km migration is higher but within the uncertainty of evidence from a field study that found pueruli of J. edwardsii were consuming 3.1 mg of lipid, with a metabolic equivalent of 113 ^ 36 J of energy for a 20-km journey (Jeffs et al. 1999). A subsequent study of a large sample of pueruli (360) caught at various distances offshore estimated that 16.5% of pueruli had insufficient energy reserves to

q 2011 by the Association for the Sciences of Limnology and Oceanography, Inc. / e-ISSN 2157-3698

Downloaded at ASLO on January 1, 2012

172

Limnology and Oceanography: Fluids and Environments

1 (2011)

250

150 113 J

100

2 da ys

50

0

0

10 20 Distance traveled (km)

4 da ys

200

20 cm s–1 15 cm s–1 7.5 cm s–1 30

which are in the range of swimming speeds for pueruli of this species (Jeffs and Holland 2000). [35] If there were no period of daytime inactivity, then the swimming model suggests a puerulus could execute a 20-km migration by continuous swimming at 5 cm s21 for a total energy demand of 90 J (at 4.5 J km21; Fig. 7). We do not know whether this behavior is an option a puerulus could elect, if its energy reserves were critically low, but it does show that a total effort less than the 113 J estimated by Jeffs et al. (1999) is achievable.
Uncertainty Analysis [36] It was shown above that relatively large variations (^30%) in the force parameters have a modest influence on swimming velocity at fixed beat frequency (Table 1). But drag also affects the work done and energy consumed. The sensitivity of energy consumption due to varying the body drag and added mass coefficient can be projected from our model (two coincident shaded areas in Fig. 6; Table 1). But in terms of energy used versus swimming speed, the sensitivity to pleopod drag is small because this coefficient controls the propulsive thrust and increased effort is harnessed into increased speed (dotted line in Fig. 6). Since energy use appears most sensitive to CDjBody, we illustrate sensitivity of PSWIM for a ¼ 0.5 to only this parameter (Fig. 7, Table 1). There is considerable sensitivity at the higher swimming speeds that we have argued are unlikely to be sustainable for a full migration, but at lower speeds the results are robust to this parameter variation. Varying the drag alters the speed at which energy consumption is at a minimum (Fig. 7, Table 1).

Fig. 8 The total energy consumed versus distance traveled when the ratio a ¼ 0.5 is assumed for the fraction of the diurnal cycle that a puerulus of Jasus edwardsii is not engaged in active swimming. Colored solid lines with symbols indicate different active swimming speeds of 7.5 (black), 15 (blue), and 20 (red) cm s21. Dashed contours indicate elapsed time (including inactivity). Green dash-dotted line and error bars show the Jeffs et al. (1999) estimate of puerulus energy available for the last 20 km of migration (113 ^ 36 J). Green shaded region indicates the energy demands of “successful migration,” that is, travel distances of 20 km or more. Yellow shaded region bracketing the 7.5 cm s21 curve indicates sensitivity to ^ 30% variation in drag coefficient of the puerulus body.

reach inshore settlement grounds (Jeffs et al. 2001a). Similar estimates of the inadequacy of the biochemical energy reserves have been reported for the puerulus of other spiny lobster species, such as the Western Australian spiny lobster, Panulirus cygnus (Phillips et al. 2006b). Using the rates of energy consumption calculated from our swimming model would suggest that a significant proportion of pueruli would have insufficient energy reserves to reach inshore settlement grounds at nighttime swimming speeds of 15 and 20 cm s21, respectively,

Table 1 Sensitivity of optimum swimming speed, Vsopt, and corresponding rate of energy consumption, P opt, to changes in parameters of swimming model that control the thrust and drag force balance, and mechanical work. CDj Body is the drag coefficient of the puerulus body (equation (7)). CDjpleopod is the drag coefficient of the pleopod cross section (equation (4)). KA is the added mass coefficient.
Parameter varied Sensitivity analysis Variation from control case V opt (cm s 21) s P opt (J km 21) Control case 0 7.49 6.68 230% 8.11 6.10 CDj Body þ 30% 6.99 7.22 230% 7.29 6.74 CDjpleopod þ 30% 6.95 6.73 230% 7.57 6.31 KA þ30% 7.05 7.00

q 2011 by the Association for the Sciences of Limnology and Oceanography, Inc. / e-ISSN 2157-3698

Total energy consumed (J)

Downloaded at ASLO on January 1, 2012

6 da ys

173

Larval lobster swimming energetics

Wilkin and Jeffs

[37] There is only modest sensitivity of total energy consumed during migration at optimum swimming speed to ^30% change in CDjBody (dotted lines bracketing the 7.5 cm s21 swimming line in Fig. 8).
Factors Influencing Migration Efficiency [38] The duration of migration is distance traveled divided by average speed, that is, R/aVS, which allows us to write the energy consumed to travel distance R (duration · PSWIM) as   F Drag PREST EðRÞ ¼ R þ aV S mM mH ð12Þ   PREST gC D V 2 S þ ; ¼R aV S mM mH

This result indicates, again, that metabolic and hydrodynamic efficiencies are unlikely to be major determinants of the maximum possible migration distance. Rather, the proportion of time a puerulus is not engaged in active swimming and the base metabolic cost that is independent of swimming are likely to be relatively more important factors in overall migration fitness.
Significance to Aquatic Environments [40] In recent years the total global landings of spiny lobster have gone into a rapid decline after remaining stable for more than 30 years (Jeffs 2010). Therefore, the decline cannot be explained by overfishing and appears to be related to marked reductions in the recruitment of postlarval stages onto coastal reefs (Caputi et al. 2010). The postlarval stage of most spiny lobster species migrates from beyond the continental shelf by actively swimming into shallow coastal waters, where they settle to become juveniles. The postlarvae do not feed during this migration but rely on stored lipid to provide sufficient energy for the entire migration. We have used a model of the hydrodynamic forces on the swimming appendages and body of the postlarvae of spiny lobsters to deduce the energy requirements for this migration. Predictions of swimming speeds and energy consumption rates derived from the model for the postlarvae of one spiny lobster species are consistent with independent measures from wild postlarvae, which provides some confidence in the outputs of the model (Jeffs et al. 1999, 2001a). Predictions by the model of the total energy requirements for postlarvae for the cross-shelf migration indicate that the biochemical energy stores previously measured in the postlarvae are marginal, with a significant proportion of postlarvae having insufficient energy stores to complete the migration. To exacerbate this situation, the addition of adverse ocean currents could impede onshore progress of the postlarvae, or poor feeding conditions caused by low productivity during the preceding lengthy larval period (16– 18 months) during which the stored energy reserves are accumulated could further limit the potential swimming range of the postlarvae. Given the apparently small margin of error for the energy stores found in spiny lobster

where we use equation (10) to express PMECH in terms of hydrodynamic efficiency and represent the drag force (equation (8)) as F Drag ¼ gC D V 2 , where g is a factor S determined by the puerulus geometry. The only parameter in equation (12) that has significant dependence on VS is CD, which varies by ^20% over the range 5 , VS , 10 cm s21. This variation is too large to wholly neglect, but if we assume for the sake of discussion that CD is constant, we can deduce an approximate expression ~ opt for the optimum swimming speed, VS , by minimizing E(R) with respect to VS:  1=3 ~ opt < PREST mM mH VS : ð13Þ 2gC D a The occurrence of the one-third power explains why the optimal swimming speeds we calculated from the full model were relatively insensitive to the value of a. Moreover, this expression reveals that the other dimensionless parameters of order 1, namely, mM and mH, will also have modest impact on the optimal swimming speed. [39] Substituting equation (13) into (12), we get ~ an approximate relation for the range R that a puerulus ~ might achieve for a given energy reserve, E, swimming opt ~ at VS :     2 mM mH 1=3 a 2=3 ~ ~ R¼E : ð14Þ 3 2gC D PREST

q 2011 by the Association for the Sciences of Limnology and Oceanography, Inc. / e-ISSN 2157-3698

Downloaded at ASLO on January 1, 2012

174

Limnology and Oceanography: Fluids and Environments

1 (2011)

postlarvae to overcome adverse migration conditions, it is possible that relatively small shifts in coastal circulation associated with climatic variability or longer-term ocean climate change have the potential to generate marked changes in the survival of postlarvae migrating to the coast. As a consequence, subsequent recruitment into coastal spiny lobster fisheries has the potential to be reduced, as has recently been observed in some commercially important spiny lobster populations (Caputi et al. 2010).
Acknowledgments Many thanks go to Natalie Managh and Vivian Ward for their illustrations, Rebecca Holland and Simon Hooker for the video recording and analyses, and Heidi Fuchs for comments on the manuscript. This research was supported by New Zealand’s Foundation for Research, Science and Technology and the Glenn Family Foundation.

References
Acosta, C. A., and M. J. Butler IV. 1999. Adaptive strategies that reduce predation on Caribbean spiny lobster postlarvae during onshore transport. Limnol. Oceanogr. 44: 494 –501, doi:10 .4319/lo.1999.44.3.0494. Blevins, R. D. 1984. Applied Fluid Dynamics Handbook. Van Nostrand Reinhold. Booth, J. D. 1994. Jasus edwardsii larval recruitment off the east coast of New Zealand. Crustaceana 66: 295 – 317, doi:10.1163 /156854094X00044. Booth, J. D., and B. Phillips. 1994. Early life history of spiny lobster. Crustaceana. 66: 271 – 294, doi:10.1163/156854094X00035. Caputi, N., R. Melville-Smith, S. de Lestang, A. Pearce, and M. Feng. 2010. The effect of climate change on the western rock lobster (Panulirus cygnus) fishery of Western Australia. Can. J. Fish. Aquat. Sci. 67: 85–96, doi:10.1139/F09-167. Chiswell, S. M., and J. D. Booth. 1999. Rock lobster Jasus edwardsii larval retention by the Wairarapa Eddy off New Zealand. Mar. Ecol. Prog. Ser. 183: 227 –240, doi:10.3354/meps183227. ´ Gracia, A., and E. Lozano. 1980. Alimentacion del bagre marino Netuma platypogon y su importancia como indicador de reclutamiento de postlarvas de longosta (Decapoda: Palinuridae), Guerrero, Mexico. An. Centro. Cienc. Mar. Limnol. Univ. Nac. Auton. Mex. 7: 199 –206. Howard, R. K. 1988. Fish predators of the western rock lobster (Panulirus cygnus George) in a nearshore nursery habitat. Aust. J. Mar. Freshwater Res. 39: 307– 316, doi:10.1071 /MF9880307. Jeffs, A. G. 2001. Can compromised condition explain early mortalities in spiny lobster culture? Pp. 64– 74. In L. H. Evans and J. B. Jones [eds.], Proceedings of the International Symposium on Lobster Health Management. Curtin University Press.

———. 2010. Status and challenges of advancing lobster aquaculture globally. J. Mar. Biol. Ass. India. 52: 320 – 326. Jeffs, A. G., S. M. Chiswell, and J. D. Booth. 2001a. Distribution and condition of pueruli of the spiny lobster Jasus edwardsii offshore from north-east New Zealand. Mar. Freshw. Res. 52: 1211 –1216, doi:10.1071/MF01182. Jeffs, A. G., and R. C. Holland. 2000. Swimming behaviour of the puerulus of the spiny lobster, Jasus edwardsii (Hutton, 1875) (Decapoda, Palinuridae). Crustaceana 73: 847– 856, doi:10 .1163/156854000504859. Jeffs, A. G., J. C. Montgomery, and C. T. Tindle. 2005. How do spiny lobster post-larvae find the coast? N. Z. J. Mar. Freshw. Res. 39: 605 – 617, doi:10.1080/00288330.2005.9517339. Jeffs, A. G., P. D. Nichols, and M. P. Bruce. 2001b. Lipid reserves used by pueruli of the spiny lobster Jasus edwardsii in crossing the continental shelf of New Zealand. Comp. Biochem. Physiol., Part A: Mol. Integr. Physiol. 129A: 305 –311, doi:10.1016 /S1095-6433(00)00348-2. Jeffs, A. G., P. D. Nichols, B. D. Mooney, K. L. Phillips, and C. F. Phleger. 2004. Identifying potential prey of the pelagic larvae of the spiny lobster Jasus edwardsii using signature lipids. Comp. Biochem. Physiol., Part B: Biochem. Mol. Biol. 137B: 487 – 507, doi:10.1016/j.cbpc.2004.02.003. Jeffs, A. G., C. F. Phleger, M. M. Nelson, B. D. Mooney, and P. D. Nichols. 2002. Marked depletion of polar fat and non-essential fatty acids following settlement by post-larvae of the spiny lobster, Jasus verreauxi. Comp. Biochem. Physiol., Part A: Mol. Integr. Physiol. 131A: 305 –311, doi:10.1016/S1095 -6433(01)00455-X. Jeffs, A. G., M. E. Willmott, and R. M. G. Wells. 1999. The use of energy stores in the puerulus of the spiny lobster Jasus edwardsii across the continental shelf of New Zealand. Comp. Biochem. Physiol., Part A: Mol. Integr. Physiol. 123A: 351 –357, doi:10.1016/S1095-6433(99)00073-2. Lemmens, J. W. T. J. 1994. Biochemical evidence for absence of feeding in puerulus larvae of the western rock lobster Panulirus cygnus (Decapoda: Palinuridae). Mar. Biol. 118: 383 –391, doi:10.1007/BF00350295. Morison, J. R., M. P. O’Brien, J. W. Johnson, and S. A. Schaaf. 1950. The force exerted by surface waves on piles. Petrol. Trans. Am. Inst. Mining Eng. 189: 149 –154. Morris, M. J., G. Gust, and J. J. Torres. 1985. Propulsion efficiency and cost of transport for copepods: a hydromechanical model of crustacean swimming. Mar. Biol. 86: 283 –295, doi:10.1007 /BF00397515. Morris, M. J., K. Kohlhage, and G. Gust. 1990. Mechanics and energetics of swimming in the small copepod Acanthocyclops robustus (Cyclopoida). Mar. Biol. 107: 83– 91, doi:10.1007 /BF01313245. Phillips, B. F., J. D. Booth, J. S. Cobb, A. G. Jeffs and P. McWilliam. 2006a. Larval and postlarval ecology. Pp. 231 – 262. In B. F. Phil-

q 2011 by the Association for the Sciences of Limnology and Oceanography, Inc. / e-ISSN 2157-3698

Downloaded at ASLO on January 1, 2012

175

Larval lobster swimming energetics

Wilkin and Jeffs

lips [ed.], Lobsters: Biology, Management, Aquaculture and Fisheries. Blackwell. Phillips, B. F., A. G. Jeffs, R. Melville-Smith, C. F. Chubb, M. M. Nelson, and P. D. Nichols. 2006b. Changes in fat and fatty acid composition of late larval and puerulus stages of the spiny lobster (Panulirus cygnus) during movement across the continental shelf of Western Australia. Comp. Biochem. Physiol., Part B: Biochem. Mol. Biol. 143: 219 –228, doi:10.1016 /j.cbpb.2005.11.009. Phleger, C. F., M. M. Nelson, B. D. Mooney, P. D. Nichols, A. J. Ritar, G. G. Smith, P. R. Hart, and A. G. Jeffs. 2001. Lipids and nutrition of the southern rock lobster, Jasus edwardsii, from hatch to puerulus. Mar. Freshw. Res. 52: 1475 –1486, doi:10 .1071/MF01071. Svetlichny, L. S., and E. S. Hubareva. 2005. The energetics of Calanus euxinus: locomotion, filtration of food and specific dynamic action. J. Plankton Res. 27: 671 – 682, doi:10.1093/plankt /fbi041.

Takahashi, Y., S. Nishida, and J. Kittaka. 1994. Histological characteristics of fat bodies in the puerulus of the rock lobster Jasus edwardsii (Hutton, 1875) (Decapoda, Palinuridae). Crustaceana 66: 318– 325, doi:10.1163/156854094X00053. Visser, A. W. 2007. Motility of zooplankton: fitness, foraging and predation. J. Plankton Res. 29: 447 –461, doi:10.1093/plankt /fbm029. Wells, R. M. G., J. Lu, A. J. R. Hickey, and A. G. Jeffs. 2001. Ontogenetic changes in enzyme activities associated with energy production in the spiny lobster, Jasus edwardsii. Comp. Biochem. Physiol., Part B: Biochem. Mol. Biol. 130: 339– 347, doi:10.1016/S1096-4959(01)00439-0.

Received: 12 March 2011 Amended: 22 June 2011 Accepted: 1 October 2011

q 2011 by the Association for the Sciences of Limnology and Oceanography, Inc. / e-ISSN 2157-3698

Downloaded at ASLO on January 1, 2012