Developing population models with data from marked individuals

Biological Conservation
Stony Brook University; University of Wisconsin-Madison; U.S. Fish and Wildlife Service
By: , and 



Population viability analysis (PVA) is a powerful tool for biodiversity assessments, but its use has been limited because of the requirements for fully specified population models such as demographic structure, density-dependence, environmental stochasticity, and specification of uncertainties. Developing a fully specified population model from commonly available data sources – notably, mark–recapture studies – remains complicated due to lack of practical methods for estimating fecundity, true survival (as opposed to apparent survival), natural temporal variability in both survival and fecundity, density-dependence in the demographic parameters, and uncertainty in model parameters. We present a general method that estimates all the key parameters required to specify a stochastic, matrix-based population model, constructed using a long-term mark–recapture dataset. Unlike standard mark–recapture analyses, our approach provides estimates of true survival rates and fecundities, their respective natural temporal variabilities, and density-dependence functions, making it possible to construct a population model for long-term projection of population dynamics. Furthermore, our method includes a formal quantification of parameter uncertainty for global (multivariate) sensitivity analysis. We apply this approach to 9 bird species and demonstrate the feasibility of using data from the Monitoring Avian Productivity and Survivorship (MAPS) program. Bias-correction factors for raw estimates of survival and fecundity derived from mark–recapture data (apparent survival and juvenile:adult ratio, respectively) were non-negligible, and corrected parameters were generally more biologically reasonable than their uncorrected counterparts. Our method allows the development of fully specified stochastic population models using a single, widely available data source, substantially reducing the barriers that have until now limited the widespread application of PVA. This method is expected to greatly enhance our understanding of the processes underlying population dynamics and our ability to analyze viability and project trends for species of conservation concern.

Publication type Article
Publication Subtype Journal Article
Title Developing population models with data from marked individuals
Series title Biological Conservation
DOI 10.1016/j.biocon.2016.02.031
Volume 197
Year Published 2016
Language English
Publisher Elsevier Science Ltd.
Publisher location Kidlington, Oxford
Contributing office(s) Upper Midwest Environmental Sciences Center
Description 10 p.
First page 190
Last page 199
Online Only (Y/N) N
Additional Online Files (Y/N) N
Google Analytic Metrics Metrics page
Additional publication details