SREL Reprint #2174

 

 

 

 

Sensitivity Analysis of Structured-Population Models for Management and Conservation

Philip Dixon, Nancy Friday, Put Ang, Selina Heppell, and Mrigesh Kshatriya

Conclusions

A common theme in this chapter has been the estimation of parameter sensitivity in structured models for population management. The choice of model structure and parameter values, while always dependent on the availability of data, is also dependent on the question to be answered; in these three cases, the models attempt to address how management should be focused on particular life stages. In the analysis of elk-and-wolf interaction, we evaluate the effect of adding age structure to a predator-prey model; although most of the simulations do not change qualitatively, we are able to analyze the potential impacts of management or wolf predation that specifically affect calf and yearling survival. Thus, the addition of age structure allows predictions for more-detailed management. In the tick section, we recast a simulation model as a matrix-transition model to evaluate analytically the sensitivity of population growth rate to parameter changes. Our elasticity analysis enables us to pinpoint the most critical time of year and life stages for tick population growth, thus providing a way to focus management efforts for population control. Finally, our comparison of stage-specific elasticities in deterministic and stochastic population models shows that in most cases, elasticity estimates do not change qualitatively in variable environments. This type of analysis is thus seen to be qualitatively invariant for populations with moderate amounts of temporal variability in the vital rates.

These sorts of calculations have two different implications for population managers. As tools for model assessment, sensitivity calculations indicate critical components of a model. Particular effort should be made to ensure that parameters with high sensitivity are accurately known and that critical parts of the model structure are appropriately specified. In this role, sensitivity and elasticity analyses are tools for model validation. As tools for management, sensitivities indicate where management actions to change vital rates have large effects on the population growth rate. In this role, sensitivity and elasticity analyses are one component among many that contribute to effective and scientifically justifiable decision making.

SREL Reprint #2174

Dixon, P., N. Friday, P. Ang, S. Heppell, and M. Kshatriya. 1997. Sensitivty analysis of structure population models for management and conservation. p. 471-513. In Structured-population models in marine, terrestrial, and freshwater systems, edited by Tuljapurkar &Caswell. International Thomson Publishing. New York.

To request a reprint

 

 
http://srel.uga.edu www.uga.edu