Dissertation - HQ

94 Vertical distribution during ontogeny
Vertical shifts cause
horizontal movement
Vertical distribution
vs. vertical migration
where there is enough light to feed but not too much to be seen by
predators becomes deeper and deeper in the course of ontogeny. Finally,
temperature also varies vertically and may affect this trade-off, because
of its influence on metabolic rates 72 : fishes staying deeper in the water
column live in colder environments, have lower metabolic rates hence
need to find less food but also grow more slowly.
Apart from influencing the probability to find food, these vertical
movements also affect how larvae are advected by currents. Indeed, a
shear is often noticeable between fast surface velocities and moderate
flow at depth, because of wind stress at the surface and/or bottom
friction at depth. Actually, most hydrographic variables (i.e. current
speed but also temperature, salinity, etc.) vary faster vertically that horizontally.
Therefore, moving vertically may have dramatic consequences,
even on horizontal advection. For example, in Chesapeake Bay, vertical
diffusion influences the result of an advection model more than horizontal
diffusion does 195 . And, indeed, vertical swimming by oyster larvae
greatly modifies their dispersal routes 196 . On coasts featuring strong
tides, synchronisation of vertical migration with tides is a very efficient
means of transport, either inshore or offshore 197,198 . Ontogenetic vertical
migration may also favour retention, as models suggest either on a
large scale (Georges Bank 84 ) or around a smaller island (Barbados 71 ).
Finally, more theoretical works suggest that exploiting vertical shear is
an efficient strategy to reach a settlement site, especially for larvae not
capable of swimming against the flow 199 .
Most of the data on vertical distribution comes from stratified sampling
by towed plankton nets 71,74,130,200–203 . An alternative for late stage
larvae are stratum-specific light traps 204 , but their limited scope restricts
their use. To understand the results of these methods correctly, it is
important to bear in mind that they describe the vertical distribution of
fish larvae and do not give direct information on their vertical migration
behaviour. If individuals move around but that the overall distribution
of the population stays the same, the range of vertical movement of each
larva would be greater than what is inferred from the distribution 79 . At
the other limit, an ontogenetic shift in distribution toward depth could
be the result of selective mortality in the surface without any movement
by larvae. Therefore caution is advised when interpreting distribution
data and trying to infer the movement of individuals, or even patches,
from it.
In this study we seek to detect and quantify ontogenetic shifts in
the vertical distribution of coral reef fish larvae, and to estimate their
impact on the advection of larvae by currents. We use repeated, large
scale, vertically stratified sampling to capture the vertical distribution
of the population of fish larvae around an oceanic island. Eventually,
an oceanographic model, calibrated by observations on the study site,
is used to advect larvae in a realistic, dynamic flow field and compare
the trajectories of passive and vertically migrating ones. Furthermore,
given the prevalence of depth stratified data and the disparateness of