Size and stage-dependent vertical migration patterns in reef-associated fish larvae off the eastern coast of Australia
Introduction
Larval fish begin life with a limited set of behavioural capabilities, but with the onset of ontogenetic development within the embryo and after hatching, they increase their competency to swim, sense and orientate over a short period of time (days to weeks; Leis, 2006). Historically, larval dispersal was thought to be driven predominately by ocean currents and the length of pelagic larval duration of the species (Williams et al., 1984). Other effects were considered negligible due to the weak swimming ability of larvae. Subsequent research has found that behaviour plays a big part in the dispersal of larval fish. Dispersal trajectories can change indirectly as current velocities change with depth (Corell et al., 2012; Paris et al., 2007), or directly through behaviours such as orientated swimming using sensory cues or predator avoidance (Fiksen et al., 2007). Larval fish can often sustain horizontal swimming speeds higher than the prevailing currents, and nearly any rate of continuous swimming (conditional on direction) allows them the potential to influence their path of dispersal (Fisher et al., 2000; Leis and Fisher, 2006; Leis, 2010). Furthermore, the larvae of many species can also detect and orientate towards settlement habitat using their senses to detect water chemistry, sound, and visual cues (Kingsford et al., 2002; Leis, 2010). These factors in combination enable larvae to choose between reefs, i.e. search for higher quality habitat or evince a preference to settle close to the natal reef (Gerlach et al., 2007; Jones et al., 2005). Larval dispersal is the main driver of population connectivity for fish with bipartite life histories, where the adult is sedentary, and therefore it is important to understand how behavioural traits can influence larval dispersal.
One commonly observed behavioural trait of fish is the tendency to migrate vertically through the water column (Boehlert et al., 1992; Leis, 1991, 1986; Rodríguez et al., 2006). The drivers behind larvae changing their vertical position, either through ontogenetic vertical migration (OVM) or diel vertical migration (DVM), could be evolutionary advantages through developmental adaptations, predator avoidance (especially visual predators in the case of diel vertical migration; Lampert, 1989), or increases in potential prey (Irisson et al., 2010; Pearre, 2003). Other more immediate drivers of change in vertical position could be changes in environmental variables (e.g. temperature) or sensed cues that can allow larvae to orientate.
The ability of fish larvae to regulate their depth is dependent on their ability to sense, orientate, and swim (behaviours that are well developed by settlement age; Leis, 2010). The two processes of vertical distribution: diel (Lampert, 1989; Sabatés, 2004) and ontogenetic (Boehlert et al., 1992; Grønkjær and Wieland, 1997) vertical migration have been documented extensively within the literature. Vertical migration can provide metabolic advantages by allowing larval fish to change their thermal environment (Wurtsbaugh and Neverman, 1988). Vertical migration can also allow variation in diet through changes in prey encountered at different depths (Llopiz and Cowen, 2009). Larvae changing their position in the water column can affect their dispersal due to depth specific current velocities (Paris et al., 2007; Paris and Cowen, 2004). Current velocities differ at depths in the water column due to processes such as Ekman transport along coasts (Sponaugle et al., 2002) and selective tidal transport in and out of estuaries (Grioche et al., 2000; Hare et al., 2005).
The general consensus is that larval reef fish trend deeper in the water column with ontogeny (Irisson et al., 2010; Paris and Cowen, 2004). Species-specific ontogenetic vertical migration patterns are relatively unknown due to the variation among and within taxonomic families, and also across different spatial contexts (Leis, 2010), leading to a requirement for localised studies. This study focuses on investigating the ontogenetic vertical migration patterns of larval reef fish off the coast of New South Wales (NSW), Australia. While there have been many studies on the vertical distribution of larval assemblages off the NSW coast sampling across inshore and offshore communities (Dempster et al., 1997), different depths (Gray and Miskiewicz, 2000), and within different water masses (Keane and Neira, 2008), changes in the vertical distribution with respect to ontogeny are less studied. OVM has been investigated for some commercially valuable species off the coast of NSW (Smith, 2003), along with more general relationships between length and depth finding contrasting trends between species (Gray, 1993). Understanding this early-life history aspect of larval fish is important for parameterising biophysical models since reducing the number of assumptions improves the accuracy of a model (Leis, 2007).
The eastern coast of Australia has complex ocean conditions, a rugged coastline with a narrow continental shelf and a strong near-shore boundary current (the East Australian Current, EAC). Previous studies have suggested that over short time periods, passive advection can explain observed dispersal patterns (Smith et al., 1999). However, vertical migration behaviour has the potential to influence dispersal in the region by allowing larval fish to avoid strong surface currents (Paris et al., 2007) such as the EAC. The EAC often spawns warm-core and cold-core eddies off the coast of New South Wales at any latitude concentrated below 32°S (Everett et al., 2012; Suthers et al., 2011). These mesoscale eddies entrain continental shelf waters, bringing with them both pelagic eggs and newly hatched larval fish and providing nutrient rich environments for them to develop and potentially reducing predation (Matis et al., 2014; Mullaney and Suthers, 2013; Shulzitski et al., 2015; Syahailatua et al., 2011). Simulation studies have shown that eddy speed, fish swimming speed (Chang et al., 2017), and the vertical distribution of fish (Condie and Condie, 2016) can influence the retention time within an eddy. Mesoscale cyclonic eddies, with their nutrient rich waters, have increased phytoplankton production (McGillicuddy et al., 1999; Waite et al., 2007), leading to increased zooplankton abundance as an eddy grows older (Govoni et al., 2010). Eddies acting as planktonic incubators (Bakun, 2006) are also dynamic with the potential for plankton displacement and abundance to change (Govoni et al., 2010), potentially driving changes in the vertical patterns of larval fish as they seek prey. Larval retention leads to a general pattern of longer fish larvae within an eddy, compared to the source coastal waters, potentially due to lower predation (Mullaney and Suthers, 2013). Depth distributions of fish larvae assemblages have been seen to have different patterns between cold-core and warm-core eddies (Muhling et al., 2007). Unless the vertical planktonic displacement is different between the coastal and eddy waters, there is little reason to expect differences in patterns of ontogenetic vertical migration, however, to our knowledge no studies have empirically investigated whether or not this is the case.
In this study, we investigated the vertical distribution of larvae in different oceanographic features (eddy versus no-eddy) off the coast of New South Wales (NSW), Australia using depth-stratified plankton tows. It must be noted that the sampling method only infers the vertical distribution and does not inform about the specific behaviour leading to these patterns (Pearre, 1979). We aimed to determine the ontogenetic patterns of vertical distribution of temperate reef fish families off NSW. These patterns were measured using the larval fish length and ontogenetic stage at the depth of sampling, allowing for comparisons of patterns between measures. We also investigated to see if the nature of these patterns is dependent on the surrounding oceanography (entrained in an eddy).
Section snippets
Materials and methods
Sampling was undertaken along the eastern coast of Australia aboard the Australian Marine National Facility's RV Investigator (2–18 June 2015, voyage number IN2015_V03). Sixteen stations were sampled across two coastal regions and two cyclonic cold core eddies (one frontal ~7 days old and one mesoscale ~26 days old; Roughan et al., 2017) in the Eastern Australian Current (EAC; Fig. 1). The location of the eddies was calculated onboard using real-time SST data, eddy centres were estimated by
Results
The seven reef-associated families comprised 2827 fish in total (a subset of 41346 identified larval fish), of which 44.3% came from the family Labridae, the only family to be sampled across all of the tows (Table 2). The surface tow at station 8 was not included in the analysis because there were no flowmeter readings to standardise the abundance, and at station 16 the surface nets were not deployed due to rough conditions (sample sizes summarised in Table 2). Labridae were the most abundant
Discussion
This study provides evidence of downward shifts of vertical distribution with ontogeny in many typical temperate reef fish families of south-eastern Australia. The mesoscale oceanographic features in general did not seem to affect the vertical distribution patterns, eddies in general contained longer larvae than the coastal waters. This study also highlights how associated environmental characteristics appear to be weak predictors of the abundance and length of the families considered, with
Conclusions
Ontogenetic vertical distributions are not uniform for reef-associated fish families of NSW, but the general pattern for these seven families was to move downward with larval development. When larval fish do exhibit patterns that could be associated with ontogenetic vertical migration, variables of length and abundance can show contrasting patterns because of the large influence of the relative abundance of different stages, suggesting capturing the abundance of ontogenetic stages at depth is a
Funding
This research was conducted with support from an Australian Government Research Training Program Scholarship.
Declaration of competing interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: None
Acknowledgements
We would like to acknowledge the effort of the captain, crew, Australian Marine National Facility staff and Chief Scientist Iain Suthers on board the RV Investigator for assistance with the running of the voyage. In addition, to Chris Howdon for help with the statistical analysis. The lead author would also like to acknowledge the examining reviewers of his thesis for improving this chapter and providing insightful discussion. This work was conducted under UNSW Animal Care & Ethics Committee
References (81)
- et al.
Vertical structure of larval fish assemblages during diel cycles in summer and winter in the southern part of Bahia de La Paz, Mexico
Estuar. Coast Shelf Sci.
(2008) - et al.
Horizontal and vertical distributions of larval fishes around an isolated oceanic island in the tropical Pacific
Deep-Sea Res. Part A Oceanogr. Res. Pap.
(1992) - et al.
Beyond hydrography: daily ichthyoplankton variability and short term oceanographic events on the Sydney continental shelf
Continent. Shelf Res.
(1997) - et al.
Larval fish assemblages in south-east Australian coastal waters: seasonal and spatial structure
Estuar. Coast Shelf Sci.
(2000) - et al.
Coral reef fish larvae settle close to home
Curr. Biol.
(2005) Are larvae of demersal fishes plankton or nekton?
Adv. Mar. Biol.
(2006)- et al.
The effects of haloclines on the vertical distribution and migration of zooplankton
J. Exp. Mar. Biol. Ecol.
(2002) - et al.
Late pelagic-stage goatfishes: distribution patterns and inferences on schooling behaviour
J. Exp. Mar. Biol. Ecol.
(1993) - et al.
Ichthyoplankton assemblage structure in two meso-scale leeuwin current eddies, eastern Indian ocean. Deep Sea research Part II: topical studies in oceanography
The Leeuwin Current and its Eddies
(2007) - et al.
Role of mesoscale eddies in shaping the spatial distribution of the coexisting Engraulis encrasicolus and Sardinella aurita larvae in the northwestern Mediterranean
J. Mar. Syst.
(2013)