10. Using Matrix Models to Focus Research and Management Efforts
157
ment efforts directed at less sensitive life stages (Groom and Pascual 1998).
Uncertainty in parameter estimates and annual variability or nonlinear changes in
vital rates following a perturbation are just a few of the real-world factors that can
alter the predicted effects of an elasticity analysis. However, it is a valuable tool
for initial management recommendations for poorly known species and as a first
step in population viability analysis (Ferriere et al. 1996).
Case Studies
We illustrate different aspects of the use of matrix projection models for aiding
management decisions in three case studies, presented here in increasing order of
complexity:
• a simple, stage-based matrix model and simulations to select between widely
divergent management alternatives for threatened sea turtles, including using a
modified age-based model to predict transient changes in population trajectories produced by shifts in the age structure of the population (Crouse et al. 1987;
Crowder et al. 1994)
• a more complex stage-based matrix, in which an individual’s life history stage
is not dependent on age, designed to select from an array of management
alternatives with multiple effects on endangered Red-cockaded Woodpeckers
(Heppell et al. 1994)
• matrix models incorporating density dependence to depict and evaluate declining Brook Trout population dynamics more accurately with respect to a variety
of management options (Marschall and Crowder 1995, 1996).
Case Study 1: Southeastern U.S. Loggerhead Sea Turtles
The evaluation of two divergent management strategies for federally listed, threatened Loggerhead Sea Turtles (Caretta caretta) in the southeastern United States
provides a case study of the use of stage-based matrix models. By the mid-1980s,
beach monitoring and nest protection projects existed on a number of beaches
throughout the southeast, staffed primarily by local volunteers. In addition to
increasing hatch success and learning about adult female fecundity and internesting behavior, these projects documented three things: (1) adult nesting females
were declining (Richardson 1982; Frazer 1983b, 1987); (2) nest, egg, and hatchling losses to erosion and both natural and feral predators were high, sometimes
approaching 95% (Hopkins and Murphy 1983); and (3) carcasses of drowned
adult and juvenile turtles of both sexes often washed ashore (i.e., “strandings”)
(Schroeder 1987).
Incidental entrapment and drowning of sea turtles was documented in a variety
of fishing gear, particularly shrimp trawls (Crouse 1984; Henwood and Stuntz
1987; Murphy and Hopkins-Murphy 1989). To address this problem, the National
Marine Fisheries Service (NMFS) developed a turtle excluder device (TED) that
released 97% of the turtles captured while allowing most shrimp to be retained in
157
ment efforts directed at less sensitive life stages (Groom and Pascual 1998).
Uncertainty in parameter estimates and annual variability or nonlinear changes in
vital rates following a perturbation are just a few of the real-world factors that can
alter the predicted effects of an elasticity analysis. However, it is a valuable tool
for initial management recommendations for poorly known species and as a first
step in population viability analysis (Ferriere et al. 1996).
Case Studies
We illustrate different aspects of the use of matrix projection models for aiding
management decisions in three case studies, presented here in increasing order of
complexity:
• a simple, stage-based matrix model and simulations to select between widely
divergent management alternatives for threatened sea turtles, including using a
modified age-based model to predict transient changes in population trajectories produced by shifts in the age structure of the population (Crouse et al. 1987;
Crowder et al. 1994)
• a more complex stage-based matrix, in which an individual’s life history stage
is not dependent on age, designed to select from an array of management
alternatives with multiple effects on endangered Red-cockaded Woodpeckers
(Heppell et al. 1994)
• matrix models incorporating density dependence to depict and evaluate declining Brook Trout population dynamics more accurately with respect to a variety
of management options (Marschall and Crowder 1995, 1996).
Case Study 1: Southeastern U.S. Loggerhead Sea Turtles
The evaluation of two divergent management strategies for federally listed, threatened Loggerhead Sea Turtles (Caretta caretta) in the southeastern United States
provides a case study of the use of stage-based matrix models. By the mid-1980s,
beach monitoring and nest protection projects existed on a number of beaches
throughout the southeast, staffed primarily by local volunteers. In addition to
increasing hatch success and learning about adult female fecundity and internesting behavior, these projects documented three things: (1) adult nesting females
were declining (Richardson 1982; Frazer 1983b, 1987); (2) nest, egg, and hatchling losses to erosion and both natural and feral predators were high, sometimes
approaching 95% (Hopkins and Murphy 1983); and (3) carcasses of drowned
adult and juvenile turtles of both sexes often washed ashore (i.e., “strandings”)
(Schroeder 1987).
Incidental entrapment and drowning of sea turtles was documented in a variety
of fishing gear, particularly shrimp trawls (Crouse 1984; Henwood and Stuntz
1987; Murphy and Hopkins-Murphy 1989). To address this problem, the National
Marine Fisheries Service (NMFS) developed a turtle excluder device (TED) that
released 97% of the turtles captured while allowing most shrimp to be retained in