THE EFFECT OF STOCHASTIC TECHNIQUE ON ESTIMATES OF POPULATION VIABILITY FROM TRANSITION MATRIX MODELS

Population viability analysis is an important tool for conservation biologists, and matrix models that incorporate stochasticity are commonly used for this purpose. However, stochastic simulations may require assumptions about the distribution of matrix parameters, and modelers often select a statistical distribution that seems reasonable without sufficient data to test its fit. We used data from long-term (5–10 year) studies with 27 populations of five perennial plant species to compare seven methods of incorporating environmental stochasticity. We estimated stochastic population growth rate (a measure of viability) using a matrix-selection method, in which whole observed matrices were selected at random at each time step of the model. In addition, we drew matrix elements (transition probabilities) at random using various statistical distributions: beta, truncated-gamma, truncated-normal, triangular, uniform, or discontinuous/observed. Recruitment rates were held constant at their observed mean values. Two methods of constraining stage-specific survival to ≤100% were also compared. Different methods of incorporating stochasticity and constraining matrix column sums interacted in their effects and resulted in different estimates of stochastic growth rate (differing by up to 16%). Modelers should be aware that when constraining stage-specific survival to 100%, different methods may introduce different levels of bias in transition element means, and when this happens, different distributions for generating random transition elements may result in different viability estimates. There was no species effect on the results and the growth rates derived from all methods were highly correlated with one another. We conclude that the absolute value of population viability estimates is sensitive to model assumptions, but the relative ranking of populations (and management treatments) is robust. Furthermore, these results are applicable to a range of perennial plants and possibly other life histories. Corresponding Editor: A. R. Solow.

[1]  Simon P. Blomberg,et al.  POPULATION DYNAMICS AND SURVIVAL OF AN ENDANGERED WALLABY: A COMPARISON OF FOUR METHODS , 2000 .

[2]  M. Nakaoka.  Demography of the marine bivalve Yoldia notabilis in fluctuating environments: an analysis using a stochastic matrix model , 1997 .

[3]  S. Ferson,et al.  Correlations, dependency bounds and extinction risks , 1995 .

[4]  C N Haas.  On modeling correlated random variables in risk assessment. , 1999, Risk analysis : an official publication of the Society for Risk Analysis.

[5]  Peter Kareiva,et al.  Modeling Population Viability for the Desert Tortoise in the Western Mojave Desert , 1994 .

[6]  E O Voit,et al.  Random Number Generation from Right‐Skewed, Symmetric, and Left‐Skewed Distributions , 2000, Risk analysis : an official publication of the Society for Risk Analysis.

[7]  R. Lande,et al.  Population Viability Analysis of Spring Chinook Salmon in the South Umpqua River, Oregon , 1997 .

[8]  L. Botsford,et al.  PROJECTING VIABILITY OF TOTOABA MACDONALDI, A POPULATION WITH UNKNOWN AGE‐DEPENDENT VARIABILITY , 1997 .

[9]  David B. Lindenmayer,et al.  A Review of the Generic Computer Programs ALEX, RAMAS/space and VORTEX for Modelling the Viability of Wildlife Metapopulations , 1995 .

[10]  Kevin Gross,et al.  Modeling Controlled Burning and Trampling Reduction for Conservation of Hudsonia montana , 1998 .

[11]  Daniel Wartenberg,et al.  Correlated Inputs in Quantitative Risk Assessment: The Effects of Distributional Shape , 1995 .

[12]  M. Evans,et al.  Statistical Distributions, Third Edition , 2001 .

[13]  H. Akçakaya,et al.  Assessing human impact despite uncertainty:viability of the northern spotted owl metapopulation in the northwestern USA , 1998, Biodiversity and Conservation.

[14]  J. Kauffman,et al.  THE EFFECT OF FIRE ON THE POPULATION VIABILITY OF AN ENDANGERED PRAIRIE PLANT , 2001 .

[15]  M. Evans.  Statistical Distributions , 2000 .

[16]  T. Kaye.  Population viability analysis of endangered plant species : an evaluation of stochastic methods and an application to a rare prairie plant , 2001 .

[17]  Mark A. Burgman,et al.  Differences and Congruencies between PVA Packages: the Importance of Sex Ratio for Predictions of Extinction Risk , 2000 .

[18]  Kevin S. McKelvey,et al.  Management of the spotted owl : A case history in conservation biology , 1996 .

[19]  A Sorribas,et al.  Estimating age-related trends in cross-sectional studies using S-distributions. , 2000, Statistics in medicine.

[20]  Ray Hilborn,et al.  The Influence of Model Structure on Conclusions about the Viability and Harvesting of Serengeti Wildebeest , 1997 .

[21]  N. Slade,et al.  The effect of skewed distributions of vital statistics on growth of age-structured populations , 1984 .

[22]  H. Caswell.  PROSPECTIVE AND RETROSPECTIVE PERTURBATION ANALYSES: THEIR ROLES IN CONSERVATION BIOLOGY , 2000 .

[23]  P. Holgate,et al.  Matrix Population Models. , 1990 .

[24]  J. Thompson.  Postdispersal Seed Predation in Lomatium Spp. (Umbelliferae): Variation among Individuals and Species , 1985 .

[25]  J. Oostermeijer,et al.  Temporal and spatial variation in the demography of Gentiana pneumonanthe, a rare perennial herb , 1996 .

[26]  Paulette Bierzychudek,et al.  The Demography of Jack‐in‐the‐Pulpit, a Forest Perennial that Changes Sex , 1982 .

[27]  Ingrid M. Parker,et al.  Evaluating approaches to the conservation of rare and endangered plants , 1994 .

[28]  H. Caswell.  Analysis of life table response experiments I. Decomposition of effects on population growth rate , 1989 .

[29]  H. Caswell,et al.  The Relative 'Importance' of Life-History Stages to Population Growth: Prospective and Retrospective Analyses , 1997 .

[30]  P B Bedient,et al.  On the effect of probability distributions of input variables in public health risk assessment. , 1997, Risk analysis : an official publication of the Society for Risk Analysis.

[31]  R. Tamura,et al.  A stabilized moment estimator for the beta-binomial distribution. , 1987, Biometrics.

[32]  Michael J. Wisdom,et al.  Factors Leading to Different Viability Predictions for a Grizzly Bear Data Set , 1996 .

[33]  Robert J. Frye,et al.  Demography and Population Viability of an Endangered Plant Species before and after Protection from Trampling , 1997 .

[34]  R. C. Hennemuth,et al.  Modelling and analysis of recruitment distributions , 1995, Environmental and Ecological Statistics.

[35]  Griffiths Da.  Maximum likelihood estimation for the beta-binomial distribution and an application to the household distribution of the total number of cases of a disease. , 1973 .

[36]  Amy W. Ando,et al.  On the Use of Demographic Models of Population Viability in Endangered Species Management , 1998 .

[37]  Carol C. Horvitz,et al.  Spatiotemporal Variation in Demographic Transitions of a Tropical Understory Herb: Projection Matrix Analysis , 1995 .

[38]  G. Wilhere,et al.  Population Viability Analysis for Red-Cockaded Woodpeckers in the Georgia Piedmont , 1995 .

[39]  Stephen P. Ellner,et al.  When is it meaningful to estimate an extinction probability , 2000 .

[40]  Shripad Tuljapurkar,et al.  Population Dynamics in Variable Environments , 1990 .

[41]  E. Dudewicz,et al.  Fitting Statistical Distributions , 2000 .

[42]  Richard Frankham,et al.  Comparison of the population viability analysis packages GAPPS, INMAT, RAMAS and VORTEX for the whooping crane (Grus americana) , 1999 .

[43]  Scott Ferson,et al.  Risk assessment in conservation biology , 1993 .

[44]  D. Ludwig.  Is it meaningful to estimate a probability of extinction , 1999 .

[45]  Bruce E. Kendall,et al.  ESTIMATING THE MAGNITUDE OF ENVIRONMENTAL STOCHASTICITY IN SURVIVORSHIP DATA , 1998 .

[46]  Charles N. Haas,et al.  Importance of Distributional Form in Characterizing Inputs to Monte Carlo Risk Assessments , 1997 .

[47]  S. Ellner,et al.  Stochastic matrix models for conservation and management: A comparative review of methods , 2001 .

[48]  E. Menges,et al.  Population viability analyses in plants: challenges and opportunities. , 2000, Trends in ecology & evolution.

[49]  M. Burgman,et al.  A stage-structured, stochastic population model for the giant kelpMacrocystis pyrifera , 1990 .

[50]  A. Raftery,et al.  Discharge Rates of Medicare Stroke Patients to Skilled Nursing Facilities: Bayesian Logistic Regression with Unobserved Heterogeneity , 1996 .

[51]  E. Witkowski,et al.  A simple population viability analysis of the critically endangered Euphorbia clivicola R.A. Dyer under four management scenarios. , 2000 .

[52]  H. Resit Akçakaya,et al.  Predictive accuracy of population viability analysis in conservation biology , 2000, Nature.

[53]  E. Menges.  Stochastic Modeling of Extinction in Plant Populations , 1992 .

[54]  K. MacCrimmon,et al.  An Analytical Study of the PERT Assumptions , 1964 .

[55]  H. Christopher Frey,et al.  Methods for Characterizing Variability and Uncertainty: Comparison of Bootstrap Simulation and Likelihood‐Based Approaches , 1999 .

[56]  H. Caswell,et al.  Stochastic demography and conservation of an endangered perennial plant (Lomatium bradshawii) in a dynamic fire regime , 2001 .