The sea urchin Diadema antillarum is a keystone herbivore in the Caribbean and its functional
extinction due to an epidemic in 1983 had a marked effect on coral reef health. Recovery of D.
antillarum has been associated with improvements in reef health but has been unexpectedly slow and
patchy with many populations persisting at low abundance on the reefs. This thesis investigates
possible reasons for low and variable recovery of D. antillarum on the reef using combined
techniques of population dynamic modelling, meta-analysis and field experimentation.
Population dynamic modelling was used to explore the implications of alternative process hypotheses
for recovery dynamics. Depensatory density dependence (Allee effects) or cultivation effects are
predicted to lead to complex dynamics characterized by switching between alternative, high and low
abundance states. Density independent variation in vital rates, e.g. due to environmental change, are
predicted to affect recovery abundance and rates without leading to complex dynamics.
Time series data explored through meta-analysis, and local recovery patterns examined in Curacao,
exhibit recovery dynamics and spatial distributions which indicate that variation in density
independent factors, rather than depensatory density dependent processes are the key drivers of
population dynamics. Recovering populations and patches of high D. antillarum abundance are both
associated with locations which exhibit high resource availability, high macroalgal abundance, few
herbivorous fish competitors, reduced predation risk and local larval retention. Specifically, I find
evidence to suggest that rates of recovery are determined by resource availability and population sizes
are controlled by predation pressure and larval retention. Interestingly, the environmental
characteristics which appear favourable to D. antillarum are also characteristics which tend to result
from high levels of human threat.
My results demonstrate that D. antillarum populations favour specific habitat characteristics which
occur in a variety of habitat types and indicate that coral reefs, upon which most attention has been
focused, actually represent relatively poor habitat for the species.
In light of this, I predict that the widespread reappearance of D. antillarum on coral reefs will only
occur once population densities in preferred, sheltered, shallow, nutrient enriched, often humanimpacted
habitats become high enough to depress habitat quality in these locations, such that forereefs
become suitable as alternative habitats.