Chapter Three: Orientation and food search behaviour of a deep sea lobster in turbulent versus laminar odour plumes
New Zealand scampi (Metanephrops challengeri) is a commercially valuable lobster species that lives in depths of 200-600 m around much of the New Zealand coast (Bell, et al., 2013; Ministry for Primary Industries, 2013). Scampi are the target of a deep sea trawl fishery which has an annual catch quota of 1191 t and generates an estimated NZ$11 million per year of sales (Ministry for Primary Industries, 2013; Tuck, 2015). Deep sea bottom trawling for scampi damages the sea floor, has high levels of bycatch and uses large quantities of fuel (Anderson, 2012; Cryer, et al., 2002; Leocádio, et al., 2012). Pots (also known as creels) are used in northern hemisphere fisheries to target the ecologically similar Norway lobster (Nephrops norvegicus) (Bell, et al., 2013; Ungfors, et al., 2013). The proposed development of a potting fishery for scampi, replicating the northern hemisphere pot fishery, would result in a more sustainable method for harvesting scampi by reducing the collateral environmental impacts and improving the fuel efficiency (Major, et al., 2017b; Suuronen, et al., 2012). In addition, live scampi landed by using potting methods can be sold into higher-value markets for live seafood (Leocádio, et al., 2012; Morello, et al., 2009; Ungfors, et al., 2013). Effective baits need to be identified for a potting fishery to be economic (Miller, 1990). Initial laboratory experiments to identify superior baits for scampi found behavioural responses were inconsistent among candidate baits (Major et al., 2017a).
Scampi, like other lobsters, search for food by tracking attractant chemicals that are released from the food source and transported to the animal as plumes of chemical odour by the localised hydrodynamics (Major, et al., 2017a; Zimmer-Faust & Case, 1982; Zimmer & Butman, 2000). Diffusion disperses chemicals in fluids that are not moving, which creates a concentration gradient away from the source (Vickers, 2000). Advection is the dominant force governing the transportation of the odour chemicals in moving fluids. The variations in the velocity of the moving fluids create small-scale hydrodynamic features, such as eddies, which are collectively known as turbulence (Vickers, 2000; Webster & Weissburg, 2009). The complexity of the odour plume is affected by the characteristics of the turbulence of the fluid movement (Webster & Weissburg, 2001). Consequently, particles of different molecular weights can be distributed throughout the plume in similar concentrations and ratios to what was originally released from the bait (Vickers, 2000). This creates filaments of the attractant chemicals in high concentrations interspersed with areas without attractants, forming a three-dimensional distribution of odorant chemicals (Atema, 1996; Moore & Crimaldi, 2004; Zimmer & Butman, 2000). The spatial variations in odour concentrations are more pronounced in higher turbulence conditions and there are shorter periods between the passage of high concentration filaments of chemicals past a fixed point (known as intermittency), when compared to low turbulence (Mead, 2003; Vickers, 2000).
In order to navigate the odour landscapes that they live in, crustaceans use a range of orientation strategies which depend on the structure of odour plumes (Weissburg & Zimmer-Faust, 1994), the hydrodynamics of their environments (Jackson, et al., 2007), the morphology of their sensory systems (Keller, et al., 2003), and their locomotory abilities (Vasey, et al., 2015). The simplest of these orientation strategies is odour-gated rheotaxis, which is when an animal moves upstream after being stimulated by the attractant chemical (Webster & Weissburg, 2009). This has been suggested to be the primary orientation strategy for blue crabs (Callinectes sapidus) and is used in combination with spatial comparisons of the chemical signals (chemo-tropotaxis) to maintain contact with the plume and progress toward the source (Weissburg & Dusenbery, 2002). In contrast, lobsters have been suggested to use a form of eddy-chemotaxis, simultaneously employing the chemosensors and mechanoreceptors on the antennules to make spatial and temporal comparisons of eddies of odorant chemicals (Atema, 1996; Moore, et al., 1991b; Pravin & Reidenbach, 2013). As turbulence affects the spatial complexity of odour plumes and the intermittency that crustaceans encounter the filaments of odorant chemicals in the plume, it has a significant effect on the foraging behaviour of a number of crustacean species, which are tuned to the turbulence they encounter in their natural habitat (Keller et al., 2001; Moore, et al., 2015; Moore & Grills, 1999).
Scampi are similar to other endobenthic crustaceans that either burrow or bury themselves in the sediment and must emerge from their burrows in order to search for food (Bell, et al., 2013; Katoh, et al., 2013). Emergence behaviour in Norway lobster and scampi has been investigated through variations in catch rates (Aguzzi, et al., 2003; Bell et al., 2008; Tuck, 2010), as the lobsters avoid capture by benthic trawls when they are either inside or at the entrance of their burrows versus emergent and foraging on the open seabed (Chapman & Rice, 1971). The emergence patterns of Norway lobsters are typically driven by the diel cycle. In shallow areas on the continental shelf (0-200 m )Norway lobster emerge nocturnally with crepuscular peaks, and in deeper areas, on the continental slope (400 m), emergence patterns are weakly diurnal (Aguzzi et al., 2009; Aguzzi, et al., 2003). Other studies have observed Norway lobster catch rates to vary in relation to tidal state (Bell, et al., 2008). Similarly, the emergence patterns of scampi, which only live on the continental slope (> 200 m), have been observed to peak at dawn and potentially during periods of higher tidal flow in both tagging studies and investigations of catch rate variation (Tuck, 2010; Tuck, et al., 2015). As scampi may be foraging for food during periods of change in tidal flows, which generates turbulence at the seafloor (Kawanisi & Yokosi, 1994; Nimmo Smith et al., 1999), there is the potential that the chemosensory systems and orientation strategies of scampi may be tuned to turbulent rather than laminar flow.
Therefore, the aim of this research is to improve our understanding of the behavioural response of scampi to odours from two types of bait (mackerel and mussel) in both turbulent and laminar flow regimes in an experimental seawater flume in the laboratory.
A total of 100 scampi were obtained from a depth of 300 m on the Chatham Rise, 250 km off the east coast of New Zealand (42-43°S, 176-177°E) in July 2015 using a short duration, 2 hour bottom trawl at slow speed of 2.8 km h-1. Scampi in good condition upon landing were transferred into aquaria with seawater adjusted to their ambient temperature at point of capture (10°C). The scampi were transported to the laboratory at the Cawthron Aquaculture Park in Nelson, New Zealand, where they were held in a recirculating aquaculture system at 10.5°C with a salinity of 36 ppt. The scampi were held in individual enclosures in plastic tanks under red light (λ > 600 nm) for at least a week to acclimatise to the system prior to commencing experiments. During acclimatisation the scampi were fed every three days with squid. Food was withheld from the scampi for 7 days before they were used in behavioural experiments to ensure they were responsive to food odour cues.
The flume was 1.5 × 0.5 × 0.3 m (L× W × D), and supplied with 10 µm and carbon filtered seawater at 10.5°C, pumped into the manifold at 10 l min-1, flowing at 1 cm s-1 and passed through a corflute collimator before reaching the behavioural arena (Fig. 3.1). The experimental arena was a 1 m long section of the flume containing seawater 30 cm deep, which extended from the end of the collimator to a weir at the opposing end of the flume that the seawater flowed over and discharged through an outlet. Individual scampi were transferred in a darkened container from their holding tanks to the experimental flume nearby. The scampi was then placed at the end of the flume arena next to the outflow weir and allowed to move around the entire arena for a 30 min acclimatisation period. The scampi were then gently ushered back to the starting point immediately in front of the outflow weir using a mesh gate to ensure the scampi were in the correct position for the experiment to start. Five grams of defrosted bait material was placed in a polyvinyl chloride (PVC) mesh bag and suspended at the antennule height of the scampi, 2 cm above the floor of the tank. If the scampi displayed any stress-related behaviour, such as tail flicking, during the transport or acclimatisation period the scampi was replaced with another animal and the acclimatisation process was repeated.
Once the bait was in position a further 30 sec was allowed for an odour plume to develop in the flume, then the mesh gate was carefully removed so as not to disrupt the plume or scampi and the experiment was allowed to run for 30 min. The experiments were conducted under infrared light and filmed from an overhead position using a Brinno TLC1200 time-lapse camera in ASAP mode. The flume was completely emptied and thoroughly cleaned between experiments.
Two baits were tested in the flume, the gonad of green-lipped mussel (Perna canaliculus) and tissue of New Zealand jack mackerel (Trachurus declivis). These baits were chosen because previous experiments had observed that scampi responded to the mackerel bait faster during the detection period than to the mussel gonad indicating differences in their chemical attractiveness to scampi (Major, et al., 2017a). Fifteen replicates were conducted for each combination of bait type (mussel versus mackerel) and flow regime (turbulent versus laminar) for a total of 60 experiments. Captive male and female scampi were randomly selected for use in the experiments from the 100 scampi available, however, given the limited supply of scampi maintaining an exact 50:50 sex ratio for the experiment was difficult and therefore a total of 28 females and 32 males were used. No gravid females were used and any scampi that had recently moulted were excluded from experiments. The orbital carapace length (OCL) of each scampi was measured (mm) after the experiment, and no scampi were subjected to repeated experiments.
The two contrasting water flow regimes, “turbulent” and “laminar” were created in the flume by altering the seawater inflow arrangements. The usages of these terms are not as formal fluid dynamic descriptors, and are used to be able to distinguish clearly between the two flow treatments. Turbulent flow in the flume was generated by passing water entering the flume through only one collimator. Laminar flow was generated by passing the incoming water through two collimators arranged in series. The flow rates of seawater were consistent for both treatments, i.e., 10 l min -1. The varying flow fields were visualised in the flume by releasing food dye from a hypodermic needle at the bait position supplied by a peristaltic pump (Fig. 3.2). In the laminar flow the plume that was released was in a consistent stream and not broken into filaments. In the close-up overhead image of the plume at the point of release (Fig. 3.2A) the small pulses of dye from the peristaltic pump can be seen, and these are also observed in the overhead image of the entire flume photo (Fig. 3.2C) as the stream of dye moves along the flume. In contrast the close up image of dye plume in the turbulent flow (Fig.3.2B) shows how the eddies and structures in the flow break the dye stream into a number of small filaments that are interspersed with clear water, a pattern that continued to develop across the entire length of the flume (Fig. 3.2D). Flows were tested once per day prior to experiments.
Behavioural phase analyses
Aspects of chemically-mediated food search behaviour have been categorised into a number of behavioural phases that can be quantified (Lee & Meyers, 1996). These phases of food search behaviour have been adapted and used in scampi (Major, et al., 2017a), and consist of:
Time to detection – the time taken from when the mesh gate enclosing the scampi is removed until detection of the bait odour by the scampi. Indicated by a marked increase in the movement of appendages that contain the chemosensitive sensilla, including flicking or grooming of the antennae and antennules, beating of mouthparts and digging with, or wiping of dactyls.
Detection period – the time from the commencement of detection behaviour until the beginning of the search period. The scampi typically continue to display detection behaviour during this period.
Search period – from when the scampi starts locomotion or orientates into the water current, to the time it arrives at the bait.
Time to reach bait – time for all of the other behavioural phases combined, i.e., from the time the barrier is lifted to when the scampi reaches the bait.
These phases of behaviour were quantified by assessing the video recording of the bait-seeking behaviour of each scampi.
Orientation pathways were digitized using the TrackMate plug-in for ImageJ available in the FIJI package (Schindelin et al., 2012) for those scampi that successfully reached the bait, i.e., 14 scampi for the turbulent mackerel treatment, 12 scampi for both the laminar mackerel and turbulent mussel gonad treatments and 9 scampi for the laminar mussel gonad treatment. Digitizing was undertaken at one frame per second and x and y spatial co-ordinates were obtained for each movement of the scampi. As the scampi can orientate in a range of directions without changing its spatial location a single reference point where the cephalothorax meets the abdomen was used as the spatial reference for digitizing and calculating five orientation parameters (Fig. 3.1). These parameters were adapted from previous research (Moore & Grills, 1999; Moore, et al., 1991b; Wolf et al., 2004) and consisted of:
Walking speed – Distance travelled by the scampi to go from the point in the previous frame to the current point (one second difference between frames).
Turn angle – The difference between the bearing that the scampi walked to get to the current point and the bearing that the scampi turns to move to the next point in the subsequent frame. Hence, low turn angles indicate that the scampi were walking in a straight line.
Heading angle – The angle between a direct bearing to the bait from the scampi’s current position and the heading that it is moving in. Hence, higher heading angles indicate that the scampi is orientating further away from the bait.
Heading angle upstream – The angle between a direct bearing upstream from the scampi’s current position and the heading that it is moving in. Hence, a heading upstream of zero would indicate rheotaxis and the scampi orientating directly into the current.
Tortuosity ratio – A measure of the directness of the orientation pathway taken by the scampi from the origin to bait destination (Benhamou, 2004), that is calculated by dividing the direct distance from the origin to the bait by the total distance travelled by the scampi. Hence, the closer tortuosity ratio is to 1 the more direct the pathway is.
For each combination of baits and flow conditions log–likelihood ratios (G – test) were used to compare the proportion of the scampi successfully reaching the bait. General linear models (GLM) were used to determine the effect of the type of bait, flow regime, size and sex of the scampi on the mean time taken for each phase of chemically-mediated food search behaviour. When the flow regime was observed to have a significant effect on the phases of behaviour post-hoc t-tests with a Holm correction were used to compare effect of the flow regimes within each of the bait treatments. Data for the measures of the time taken to complete each behavioural phase were tested for normality and homoscedasticity using Shapiro’s and Levene’s tests respectively. When the raw data did not meet these assumptions, it was transformed using either natural logarithm or square-root functions.
For all five of the orientation parameters (walking speed, turn angle, heading angle, heading upstream, tortuosity ratio), a mean value was calculated for each scampi over the duration of the search period when they were actively looking for the bait, and then used in the subsequent statistical analyses (Moore & Grills, 1999). General linear models were then used to compare each of the orientation parameters in relation to the experimental treatments, i.e., type of baits, flow regime, as well as testing for any effect due to size and sex of the scampi. When the GLMs found a significant difference between the flow treatments for an orientation parameter, the means for the laminar and turbulent flow regimes within each of the bait treatments were compared using post-hoc t-tests with a Holm correction for protecting against inflated Type I error rate due to multiple comparisons. General linear models were run using the base R program (R Core Team, 2016), and multiple comparisons using the Multcomp package .
The GLMs for both the phases of chemosensory behaviour and orientation parameters were structured as so: Y = α+ 1 + 2 + 3 + 4 , with flow, sex and bait as fixed variables, while size was included as a continuous variable.
To investigate the relationship between the means and the variances of the different orientation parameters with the distance the scampi were from the bait, the means and variances of each parameter were binned into 5 cm distance intervals from the bait, and then analysed using either linear or 2nd order polynomial regression analyses. The curves and intercepts of the regressions for the two flow regimes were compared for each of the two types of bait using an analysis of covariance (ANCOVA). All means are reported with standard errors (S.E.).
Success in reaching the baits
Fourteen of the 15 (93%) scampi reached the mackerel bait and 12 of the 15 (80%) scampi reached the mussel bait within the 30 min maximum experimental period in the turbulent flow. Twelve of the 15 scampi (80%) reached the mackerel bait and 9 of the 15 scampi reached the mussel bait (60%) in the laminar flow. The different flow regimes did not alter the success rates for scampi reaching the baits, either overall (G = 5.03, P > 0.05), or for either the mackerel (G = 1.20, P > 0.05) or mussel (G = 1.45, P > 0.05) baits alone.
Table of Contents
Table of Contents
List of Figures
List of Tables
1 Chapter One: General Introduction
1.2 Ecology of New Zealand scampi
1.3 Scampi fisheries
1.4 Review of crustacean chemoattraction
1.5 Research aims and thesis structure.
2 Chapter Two: Laboratory investigations of the foraging behaviour of New Zealand scampi
2.2 Materials and methods
3 Chapter Three: Orientation and food search behaviour of a deep sea lobster in turbulent versus laminar odour plumes
3.2.1 Experimental animals
4 Chapter Four: Laboratory comparison of potential baits for potting New Zealand scampi
5 Chapter Five: Factors affecting bycatch in a developing New Zealand scampi potting fishery
6 Chapter Six: General discussion
6.2 Chemosensory behaviour of New Zealand scampi
6.3 Application to a potting fishery
6.4 Contribution to the knowledge of chemically-mediated food search behaviour in decapods
6.5 General conclusions
List of References
GET THE COMPLETE PROJECT
Chemoattraction in New Zealand scampi (Metanephrops challengeri): Identifying effective baits to develop a potting fishery