352
15. Fish Population Responses to Sea Floor Habitat A lteratio n
five differences in juvenile cod mortality, rather than absolute values. In this
way, we can comment on the trajectory of the system given a particular perturbation, without predicting absolute numbers of surviving cod.
Results of the dynamic model demonstrate distinct patterns in juvenile cod
survivorship, and linked those patterns to sea floor habitat quality and fishinginduced alteration of sea floor habitat. The model indicates that the vagility
of juvenile cod and post-settlement juvenile cod density are each significant
factors in survivorship of early benthic-phase juvenile cod, particularly in the
context of MPAs designed to protect sea floor habitat. This is true for each of
the three simulated forms of density-dependent natural mortality.
Recent modeling studies have addressed explicitly the use of MPAs for
fish conservation and fishe ries management in outer continental shelf systems (Polacheck 1990; Holland 1993; Man et al. 1995; Ruth 1995; ].S.
Nowlis, pe rsonal communication). Most of these model s deal primarily with
fish as mature , reproductive adults, with early life history often represented
in the models by pelagic dispersal of eggs and larvae. Similarly, fish mortality is mainly driven by fishing effort, with no coupling of fish popul ations to
habitat-med iated processes, or fishing effects on those processes. In our
model , juvenile cod settle to the sea floor in the first period , and settlement
is uniform to all habitat types. This is consistent with field observations of
juvenile cod in the near-shore environment (Tupper & Boutilier 1995a) and
on the northeast peak of Georges Bank (Lough et al. 1989). These studies
indicate that settlement occurs over all hab itat types, with differences in
post-settlement juven ile abundance between habitats attributed to habitatmediated, predator-induced mortality. Natural mortality is specified in the
model as habitat-specific, predato r-induced, and dens ity-depende nt.
To account for the variety of predator-prey-habitat interactions that occur
on the sea floor, the model uses density-dependent juvenile mortality to capture three functional response curves. Although each of the three individual
functional response curves may be insufficient to characterize these conditions, the comparison of model results indicates that, while actual survivorship varied, the patterns of survivorship remain similar in all three cases.
Results from the model suggest that juvenile cod survivorship (fish < 14
em total length) is highly sensitive to juvenile movement rates, particularly
with respe ct to the design of MPAs. Where no juvenile cod movement occurs, maximum survivorship within MPAs is constant for all MPA sizes
greater than zero . Such site fidelity has been observed among recently settled juveniles in areas of ade quate cover (Tupper & Boutilier 1995a). Our
model results also indicated that the siting of small MPAs is critical relative
to the location of sea floor habitat. While regional spe cies assemblage patterns in the Gulf of Maine can be defined by temp erature and depth contours (Over holtz & Tyler 1985; Gabriel 1992), fish movement at smaller
scales, relative to particular habitat types, remains highly uncertain. It is
precisely for the smaller scales in our model that movement data is required
for the design and implementation of MPAs for fish conservation and management. For larger MPAs (> 50% habitat protected), the influence of move-
15. Fish Population Responses to Sea Floor Habitat A lteratio n
five differences in juvenile cod mortality, rather than absolute values. In this
way, we can comment on the trajectory of the system given a particular perturbation, without predicting absolute numbers of surviving cod.
Results of the dynamic model demonstrate distinct patterns in juvenile cod
survivorship, and linked those patterns to sea floor habitat quality and fishinginduced alteration of sea floor habitat. The model indicates that the vagility
of juvenile cod and post-settlement juvenile cod density are each significant
factors in survivorship of early benthic-phase juvenile cod, particularly in the
context of MPAs designed to protect sea floor habitat. This is true for each of
the three simulated forms of density-dependent natural mortality.
Recent modeling studies have addressed explicitly the use of MPAs for
fish conservation and fishe ries management in outer continental shelf systems (Polacheck 1990; Holland 1993; Man et al. 1995; Ruth 1995; ].S.
Nowlis, pe rsonal communication). Most of these model s deal primarily with
fish as mature , reproductive adults, with early life history often represented
in the models by pelagic dispersal of eggs and larvae. Similarly, fish mortality is mainly driven by fishing effort, with no coupling of fish popul ations to
habitat-med iated processes, or fishing effects on those processes. In our
model , juvenile cod settle to the sea floor in the first period , and settlement
is uniform to all habitat types. This is consistent with field observations of
juvenile cod in the near-shore environment (Tupper & Boutilier 1995a) and
on the northeast peak of Georges Bank (Lough et al. 1989). These studies
indicate that settlement occurs over all hab itat types, with differences in
post-settlement juven ile abundance between habitats attributed to habitatmediated, predator-induced mortality. Natural mortality is specified in the
model as habitat-specific, predato r-induced, and dens ity-depende nt.
To account for the variety of predator-prey-habitat interactions that occur
on the sea floor, the model uses density-dependent juvenile mortality to capture three functional response curves. Although each of the three individual
functional response curves may be insufficient to characterize these conditions, the comparison of model results indicates that, while actual survivorship varied, the patterns of survivorship remain similar in all three cases.
Results from the model suggest that juvenile cod survivorship (fish < 14
em total length) is highly sensitive to juvenile movement rates, particularly
with respe ct to the design of MPAs. Where no juvenile cod movement occurs, maximum survivorship within MPAs is constant for all MPA sizes
greater than zero . Such site fidelity has been observed among recently settled juveniles in areas of ade quate cover (Tupper & Boutilier 1995a). Our
model results also indicated that the siting of small MPAs is critical relative
to the location of sea floor habitat. While regional spe cies assemblage patterns in the Gulf of Maine can be defined by temp erature and depth contours (Over holtz & Tyler 1985; Gabriel 1992), fish movement at smaller
scales, relative to particular habitat types, remains highly uncertain. It is
precisely for the smaller scales in our model that movement data is required
for the design and implementation of MPAs for fish conservation and management. For larger MPAs (> 50% habitat protected), the influence of move-