territories in winter is a more cost-effective way to reduce depredations
than reactive or population control strategies.
2.4 Lessons Learned
In planning the recovery of an endangered species, models are typically
used to estimate the likelihood of extinction and to set minimum viable
population sizes for recovery targets. However, as demonstrated by our
applications to wolf recovery, models can also be used to address various
management questions that arise during the implementation of the recovery plan. In our studies, the management questions involved predicting
the potential impacts of human-caused mortality, regional environmental
conditions (external threats), and disturbance on the persistence of wolf
populations. In addition, the management questions involved predicting the
relative performance of different strategies for controlling wolf population
size and depredation. As a result of these applications, we learned a number
of lessons about management-oriented modeling (Table 2.1). Many of these
lessons are consistent with pragmatic guidelines that have been proposed
for interdisciplinary modeling projects (Starfield 1997; Nicolson et al.
2002).
A measure of a modeling project’s success is the degree to which the
results are considered in the development of resource management policy.
We found that working in teams that included both expert biologists and
managers (Rule No. 1) and carefully defining the management questions
(Rule No. 2) was absolutely necessary to fulfill this measure of success.
When we involved expert biologists and managers in each phase of model
construction and evaluation, the simulation results comparing management
strategies and predicting relative effects of environmental factors were
credible and informative. Furthermore, by carefully delimiting the management questions, we could better decide and defend which details of
wolf demography and behavior were important to include in the model
(Rule No. 7).
Our partners understood that the purpose of our modeling exercises was
to predict the relative effects of alternative management strategies or
different environmental scenarios. Framing our simulation results in relative terms helped our teams gain insights about the management problems,
which was more useful and reliable than attempting to predict population attributes precisely under uncertain future conditions (Rule No. 3).
Thorough sensitivity analyses were then used to determine how robust the
rankings of performance or effects were to changes in uncertain parameters of wolf demography (Rule No. 11). This approach is consistent with an
emerging consensus among people involved in endangered-species management that demographic models should be used cautiously in population
viability analysis because of concerns about the accuracy of predictions
2. Modeling for Endangered-Species Recovery
39
than reactive or population control strategies.
2.4 Lessons Learned
In planning the recovery of an endangered species, models are typically
used to estimate the likelihood of extinction and to set minimum viable
population sizes for recovery targets. However, as demonstrated by our
applications to wolf recovery, models can also be used to address various
management questions that arise during the implementation of the recovery plan. In our studies, the management questions involved predicting
the potential impacts of human-caused mortality, regional environmental
conditions (external threats), and disturbance on the persistence of wolf
populations. In addition, the management questions involved predicting the
relative performance of different strategies for controlling wolf population
size and depredation. As a result of these applications, we learned a number
of lessons about management-oriented modeling (Table 2.1). Many of these
lessons are consistent with pragmatic guidelines that have been proposed
for interdisciplinary modeling projects (Starfield 1997; Nicolson et al.
2002).
A measure of a modeling project’s success is the degree to which the
results are considered in the development of resource management policy.
We found that working in teams that included both expert biologists and
managers (Rule No. 1) and carefully defining the management questions
(Rule No. 2) was absolutely necessary to fulfill this measure of success.
When we involved expert biologists and managers in each phase of model
construction and evaluation, the simulation results comparing management
strategies and predicting relative effects of environmental factors were
credible and informative. Furthermore, by carefully delimiting the management questions, we could better decide and defend which details of
wolf demography and behavior were important to include in the model
(Rule No. 7).
Our partners understood that the purpose of our modeling exercises was
to predict the relative effects of alternative management strategies or
different environmental scenarios. Framing our simulation results in relative terms helped our teams gain insights about the management problems,
which was more useful and reliable than attempting to predict population attributes precisely under uncertain future conditions (Rule No. 3).
Thorough sensitivity analyses were then used to determine how robust the
rankings of performance or effects were to changes in uncertain parameters of wolf demography (Rule No. 11). This approach is consistent with an
emerging consensus among people involved in endangered-species management that demographic models should be used cautiously in population
viability analysis because of concerns about the accuracy of predictions
2. Modeling for Endangered-Species Recovery
39