Demographic Characteristics of Extinction in a Small, Insular Population of House Sparrows in Northern Norway

Abstract:  In conservation ecology there is an urgent need for indicators that can be used to predict the risk of extinction of populations. Identifying extinction-prone populations has been difficult because few data sets on the demographic characteristics of the final stage to extinction are available and because of problems in separating out stochastic effects from changes in the expected dynamics. We documented the demographic changes that occurred during the period prior to extinction of a small island population of House Sparrows ( Passer domesticus ) after the end of permanent human settlement. A mark-recapture analysis revealed that this decline to extinction was mainly due to increased mortality after closure of the last farm that resulted in a negative long-term-specific growth rate. No change occurred in either the structural composition (breeding sex ratio and age distribution) of the population or in female recruitment. No male, however, recruits were produced on the island after the farm closure. Based on a simple, stochastic, density-dependent model we constructed a population prediction interval (PPI) to estimate the time to extinction. The 95% PPI slightly overestimated the time to extinction with large uncertainty in predictions, especially due to the influence of demographic stochasticity and parameter drift. Our results strongly emphasize the importance of access to data on temporal variation that can be used to parameterize simple population models that allow estimation of critical parameters for credible prediction of time to extinction.     

A Habitat-Based Metapopulation Model of the California Gnatcatcher

We present an analysis of the metapopulation dynamics of the federally threatened coastal California Gnatcatcher (Polioptila c. californica) for an approximately 850 km² region of Orange County, California. We developed and validated a habitat suitability model for this species using data on topography, vegetation, and locations of gnatcatcher pair observations. Using this habitat model, we calculated the spatial structure of the metapopulation, including size and location of habitat patches and the distances among them. We used data based on field studies to estimate parameters such as survival, fecundity, dispersal, and catastrophes, and combined these parameters with the spatial structure to build a stage-structured, stochastic, spatially-explicit metapopulation model. The model predicted a fast decline and high risk of population extinction with most combinations of parameters. Results were most sensitive to density-dependent effects, the probability of weather-related catastrophes, adult survival, and adult fecundity. Based on data used in the model, the greatest difference in results was given when the simulation's time horizon was only a few decades, suggesting that modeling based on longer or shorter time horizons may underestimate the effects of alternative management actions.     

Extinction times in experimental populations

Predicting population extinctions is a key element of quantitative conservation biology and population ecology. Although stochastic population theories have long been used to obtain theoretical distributions of population extinction times, model-based predictions have rarely been tested. Here I report results from a quantitative analysis of extinction time in 281 experimental populations of water fleas (Daphnia magna) in variable environments. To my knowledge, this is the first quantitative estimate of the shape of the distribution of population extinction times based on extinction data for any species. The finding that the distribution of population extinction times was extraordinarily peaked is consistent with theoretical predictions for density-independent populations, but inconsistent with predictions for density-dependent populations. The tail of the extinction time distribution was not exponential. These results imply that our current theories of extinction are inadequate. Future work should focus on how demographic stochasticity scales with population size and effects of nonrandom variable environments on population growth and decline.     

Influence of habitat quality, population size, patch size, and connectivity on patch-occupancy dynamics of the middle spotted woodpecker

Despite extensive research on the effects of habitat fragmentation, the ecological mechanisms underlying colonization and extinction processes are poorly known, but knowledge of these mechanisms is essential to understanding the distribution and persistence of populations in fragmented habitats. We examined these mechanisms through multiseason occupancy models that elucidated patch-occupancy dynamics of Middle Spotted Woodpeckers (Dendrocopos medius) in northwestern Spain. The number of occupied patches was relatively stable from 2000 to 2010 (15-24% of 101 patches occupied every year) because extinction was balanced by recolonization. Larger and higher quality patches (i.e., higher density of oaks >37 cm dbh [diameter at breast height]) were more likely to be occupied. Habitat quality (i.e., density of large oaks) explained more variation in patch colonization and extinction than did patch size and connectivity, which were both weakly associated with probabilities of turnover. Patches of higher quality were more likely to be colonized than patches of lower quality. Populations in high-quality patches were less likely to become extinct. In addition, extinction in a patch was strongly associated with local population size but not with patch size, which means the latter may not be a good surrogate of population size in assessments of extinction probability. Our results suggest that habitat quality may be a primary driver of patch-occupancy dynamics and may increase the accuracy of models of population survival. We encourage comparisons of competing models that assess occupancy, colonization, and extinction probabilities in a single analytical framework (e.g., dynamic occupancy models) so as to shed light on the association of habitat quality and patch geometry with colonization and extinction processes in different settings and species.     

Population Persistence in Florida Torreya: Comparing Modeled Projections of a Declining Coniferous Tree

Abstract: The Florida torreya (   Torreya taxifolia ) is a coniferous tree endemic to a 35-km stretch of bluffs and ravines along the east side of the Apalachicola River in northern Florida and southern Georgia. This formerly locally abundant tree has declined as a result of disease during the 1950s and is on the U.S. endangered species list. With no seed production in the wild, this species is headed toward extinction. We conducted a survey on roughly 200 trees from 1988 to 1996 and used these data to estimate the likelihood of population persistence during the next several decades. We compared a stage-class transition matrix model ( RAMAS) and an individual-based stochastic model ( TORSIM) of growth and mortality to project future populations. Given the current lack of seed production in the wild, all models predict extinction. The question of concern is the imminence of this predicted extinction. Differing predicted times to extinction would suggest different immediate management recommendations. Both models predicted an over 90% likelihood of persistence during the next 50 years. Predictions differed in that the transition matrix model was less optimistic than the individual-based model regarding persistence. Model sensitivity analysis showed that the results were robust to significant decreases in growth and sprouting probabilities. Submodels identified different persistence likelihoods in different populations. Nonetheless, unless management of the population can facilitate maturation and seed production, extinction of this species in the wild is inevitable.     

Erosion of Heterozygosity in Fluctuating Populations

Abstract: Demographic, environmental, and genetic stochasticity threaten the persistence of isolated populations. The relative importance of these intertwining factors remains unresolved, but a common view is that random demographic and environmental events will usually drive small populations to the brink of extinction before genetic deterioration poses a serious threat. To evaluate the potential importance of genetic factors, we analyzed a model linking demographic and environmental conditions to the loss of genetic diversity in isolated populations undergoing natural levels of fluctuation. Nongenetic processes—environmental stochasticity and population demography—were modeled according to a bounded diffusion process. Genetic processes were modeled by quantifying the rate of drift according to the effective population size, which was predicted from the same parameters used to describe the nongenetic processes. We combined these models to predict the heterozygosity remaining at the time of extinction, as predicted by the nongenetic portion of the model. Our model predicts that many populations will lose most or all of their neutral genetic diversity before nongenetic random events lead to extinction. Given the abundant evidence for inbreeding depression and recent evidence for elevated extinction rates of inbred populations, our findings suggest that inbreeding may be a greater general threat to population persistence than is generally recognized. Therefore, conservation biologists should not ignore the genetic component of extinction risk when assessing species endangerment and developing recovery plans.     

Forecasting extinction risk with nonstationary matrix models.

Matrix population growth models are standard tools for forecasting population change and for managing rare species, but they are less useful for predicting extinction risk in the face of changing environmental conditions. Deterministic models provide point estimates of lambda, the finite rate of increase, as well as measures of matrix sensitivity and elasticity. Stationary matrix models can be used to estimate extinction risk in a variable environment, but they assume that the matrix elements are randomly sampled from a stationary (i.e., non-changing) distribution. Here we outline a method for using nonstationary matrix models to construct realistic forecasts of population fluctuation in changing environments. Our method requires three pieces of data: (1) field estimates of transition matrix elements, (2) experimental data on the demographic responses of populations to altered environmental conditions, and (3) forecasting data on environmental drivers. These three pieces of data are combined to generate a series of sequential transition matrices that emulate a pattern of long-term change in environmental drivers. Realistic estimates of population persistence and extinction risk can be derived from stochastic permutations of such a model. We illustrate the steps of this analysis with data from two populations of Sarracenia purpurea growing in northern New England. Sarracenia purpurea is a perennial carnivorous plant that is potentially at risk of local extinction because of increased nitrogen deposition. Long-term monitoring records or models of environmental change can be used to generate time series of driver variables under different scenarios of changing environments. Both manipulative and natural experiments can be used to construct a linking function that describes how matrix parameters change as a function of the environmental driver. This synthetic modeling approach provides quantitative estimates of extinction probability that have an explicit mechanistic basis.     

Population Viability Analysis for an Endangered Plant

Abstract: Demographic modeling is used to understand the population viability of Furbish's lousewort, Pedicularis furbishiae , a perennial plant species endemic to the St. John River Valley in northern Maine. Environment-specific summaries of demographic parameters (survivorship, growth, and fecundity) over four years, organized into stage-based projection matrices, provide predictions of future population dynamics given a deterministic extension of past conditions. Stochastic modeling, using (I) empirically observed variation in demographic parameters, and (2) estimated rates of natural catastrophes, leads to predictions of extinction probability.
P. furbishiae viability has varied widely over the study period Viable populations with finite rates of increase > 1 are found where cover is low, woody plants do not dominate, and disturbance does not occur. Rates of increase vary over time, suggesting that stochastic analyses would be realistic. Stochastic measures of population viability incorporating environmental variation suggest that early successional environments, especially wetter sites, can support viable populations in the absence of disturbance. However; observed rates of natural catastrophe dominate viability estimates of individual populations. Metapopulation dynamics feature extinction rates that are greater than recolonization rates, and may be affected by land use in the watershed Species management needs to consider a large-scale view of the riverine corridor.     

Fitting complex population models by combining particle filters with Markov chain Monte Carlo

We show how a recent framework combining Markov chain Monte Carlo (MCMC) with particle filters (PFMCMC) may be used to estimate population state-space models. With the purpose of utilizing the strengths of each method, PFMCMC explores hidden states by particle filters, while process and observation parameters are estimated using an MCMC algorithm. PFMCMC is exemplified by analyzing time series data on a red kangaroo (Macropus rufus) population in New South Wales, Australia, using MCMC over model parameters based on an adaptive Metropolis-Hastings algorithm. We fit three population models to these data; a density-dependent logistic diffusion model with environmental variance, an unregulated stochastic exponential growth model, and a random-walk model. Bayes factors and posterior model probabilities show that there is little support for density dependence and that the random-walk model is the most parsimonious model. The particle filter Metropolis-Hastings algorithm is a brute-force method that may be used to fit a range of complex population models. Implementation is straightforward and less involved than standard MCMC for many models, and marginal densities for model selection can be obtained with little additional effort. The cost is mainly computational, resulting in long running times that may be improved by parallelizing the algorithm.     

Extinction Rate Estimates for Plant Populations in Revisitation Studies: Importance of Detectability

Abstract:  Many researchers have obtained extinction-rate estimates for plant populations by comparing historical and current records of occurrence. A population that is no longer found is assumed to have gone extinct. Extinction can then be related to characteristics of these populations, such as habitat type, size, or species, to test ideas about what factors may affect extinction. Such studies neglect the fact that a population may be overlooked, however, which may bias estimates of extinction rates upward. In addition, if populations are unequally detectable across groups to be compared, such as habitat type or population size, comparisons become distorted to an unknown degree. To illustrate the problem, I simulated two data sets, assuming a constant extinction rate, in which populations occurred in different habitats or habitats of different size and these factors affected their detectability. The conventional analysis implicitly assumed that detectability equalled 1 and used logistic regression to estimate extinction rates. It wrongly identified habitat and population size as factors affecting extinction risk. In contrast, with capture-recapture methods, unbiased estimates of extinction rates were recovered. I argue that capture-recapture methods should be considered more often in estimations of demographic parameters in plant populations and communities.     

The importance of space, time, and stochasticity to the demography and management of Alliaria petiolata

As population modeling is increasingly called upon to guide policy and management, it is important that we understand not only the central tendencies of our study systems, but the consequences of their variation in space and time as well. The invasive plant Alliaria petiolata (garlic mustard) is actively managed in the United States and is the focus of a developing biological control program. Two weevils (Coleoptera: Curculionidae: Ceutorhynchus) that reduce fecundity (C. alliariae) and rosette survival plus fecundity (C. scrobicollis) are under consideration for release pending host specificity testing. We used a demographic modeling approach to (1) quantify variability in A. petiolata growth and vital rates and (2) assess the potential for single- or multiple-agent biocontrol to suppress growth of 12 A. petiolata populations in Illinois and Michigan studied over three plant generations. We used perturbation analyses and simulation models with stochastic environments to estimate stochastic growth rates (lambda(S)) and predict the probability of successful management using either a single biocontrol agent or two agent species together. Not all populations exhibited invasive dynamics. Estimates of lambda(S) ranged from 0.78 to 2.21 across sites, while annual, deterministic growth (lambda) varied up to sevenfold within individual sites. Given our knowledge of the biocontrol agents, this analysis suggests that C. scrobicollis alone may control A. petiolata at up to 63% of our study sites where lambda >1, with the combination of both agents predicted to succeed at 88% of sites. Across sites and years, the elasticity rankings were dependent on lambda. Reductions of rosette survival, fecundity, or germination of new seeds are predicted to cause the greatest reduction of lambda in growing populations. In declining populations, transitions affecting seed bank survival have the greatest effect on lambda. This contrasts with past analyses that varied parameters individually in an otherwise constant matrix, which may yield unrealistic predictions by decoupling natural parameter covariances. Overall, comparisons of stochastic and deterministic growth rates illustrate how analyses of individual populations or years could misguide management or fail to characterize complex traits such as invasiveness that emerge as attributes of populations rather than species.     

Habitat Change and Plant Demography: Assessing the Extinction Risk of a Formerly Common Grassland Perennial

Abstract:  An important aim of conservation biology is to understand how habitat change affects the dynamics and extinction risk of populations. We used matrix models to analyze the effect of habitat degradation on the demography of the declining perennial plant Trifolium montanum in 9 calcareous grasslands in Germany over 4 years and experimentally tested the effect of grassland management. Finite population growth rates (λ) decreased with light competition, measured as leaf-area index above T. montanum plants. At unmanaged sites λ was <1 due to lower recruitment and lower survival and flowering probability of large plants. Nevertheless, in stochastic simulations, extinction of unmanaged populations of 100 flowering plants was delayed for several decades. Clipping as a management technique rapidly increased population growth because of higher survival and flowering probability of large plants in managed than in unmanaged plots. Transition-matrix simulations from these plots indicated grazing or mowing every second year would be sufficient to ensure a growth rate ≥1 if conditions stayed the same. At frequently grazed sites, the finite growth rate was approximately 1 in most populations of T. montanum . In stochastic simulations, the extinction risk of even relatively small grazed populations was low, but about half the extant populations of T. montanum in central Germany are smaller than would be sufficient for a probability of survival of >95% over 100 years. We conclude that habitat change after cessation of management strongly reduces recruitment and survival of established individuals of this perennial plant. Nevertheless, our results suggest extinction processes may take a long time in perennial plants, resulting in an extinction debt. Even if management is frequent, many remnant populations of T. montanum may be at risk because of their small size, but even a slight increase in size could considerably reduce their extinction risk.     

Predicting [corrected] extinction risk in spite of predator-prey oscillations.

Most population viability analyses (PVA) assume that the effects of species interactions are subsumed by population-level parameters. We examine how robust five commonly used PVA models are to violations of this assumption. We develop a stochastic, stage-structured predator-prey model and simulate prey population vital rates and abundance. We then use simulated data to parameterize and estimate risk for three demographic models (static projection matrix, stochastic projection matrix, stochastic vital rate matrix) and two time series models (diffusion approximation [DA], corrupted diffusion approximation [CDA]). Model bias is measured as the absolute deviation between estimated and observed quasi-extinction risk. Our results highlight three generalities about the application of single-species models to multi-species conservation problems. First, our collective model results suggest that most single-species PVA models overestimate extinction risk when species interactions cause periodic variation in abundance. Second, the DA model produces the most (conservatively) biased risk forecasts. Finally, the CDA model is the most robust PVA to population cycles caused by species interactions. CDA models produce virtually unbiased and relatively precise risk estimates even when populations cycle strongly. High performance of simple time series models like the CDA owes to their ability to effectively partition stochastic and deterministic sources of variation in population abundance.     

The Reliability of Using Population Viability Analysis for Risk Classification of Species

I examine whether or not it is appropriate to use extinction probabilities generated by population viability analyses, based on best estimates for model parameters, as criteria for listing species in Red Data Book categories as recently proposed by the World Conservation Union. Such extinction probabilities are influenced by how accurately model parameters are estimated and by how accurately the models depict actual population dynamics. I evaluate the effect of uncertainty in parameter estimation through simulations. Simulations based on Steller sea lions were used to evaluate bias and precision in estimates of probability of extinction and to consider the performance of two proposed classification schemes. Extinction time estimates were biased (because of violation of the assumption of stable age distribution) and underestimated the variability of probability of extinction for a given time (primarily because of uncertainty in parameter estimation). Bias and precision in extinction probabilities are important when these probabilities are used to compare the risk of extinction between species. Suggestions are given for population viability analysis techniques that incorporate parameter uncertainty. I conclude that testing classification schemes with simulations using quantitative performance objectives should precede adoption of quantitative listing criteria.     

Inference about density and temporary emigration in unmarked populations

Few species are distributed uniformly in space, and populations of mobile organisms are rarely closed with respect to movement, yet many models of density rely upon these assumptions. We present a hierarchical model allowing inference about the density of unmarked populations subject to temporary emigration and imperfect detection. The model can be fit to data collected using a variety of standard survey methods such as repeated point counts in which removal sampling, double-observer sampling, or distance sampling is used during each count. Simulation studies demonstrated that parameter estimators are unbiased when temporary emigration is either "completely random" or is determined by the size and location of home ranges relative to survey points. We also applied the model to repeated removal sampling data collected on Chestnut-sided Warblers (Dendroica pensylvancia) in the White Mountain National Forest, U.S.A. The density estimate from our model, 1.09 birds/ha, was similar to an estimate of 1.11 birds/ha produced by an intensive spot-mapping effort. Our model is also applicable when processes other than temporary emigration affect the probability of being available for detection, such as in studies using cue counts. Functions to implement the model have been added to the R package unmarked.     

Predicting Extinction Times from Environmental Stochasticity and Carrying Capacity

Managers of small populations often need to estimate the expected time to extinction T_e of their charges. Useful models for extinction times must be ecologically realistic and depend on measurable parameters. Many populations become extinct due to environmental stochasticity, even when the carrying capacity K is stable and the expected growth rate is positive. A model is proposed that gives T_e by diffusion analysis of the log population size n_t (= log_e N_t). The model population grows according to the equation N_t+1 = R_tN_t, with K as a ceiling. Application of the model requires estimation of the parameters k = logK, r_d = the expected change in n, v_r = Variance(log R), and ϱ the autocorrelation of the r_t. These are readily calculable from annual census data (r_d is trickiest to estimate). General formulas for T_e are derived. As a special case, when environmental fluctuations overwhelm expected growth (that is r_d 0), T_e = 2n_o(k - n_o/2)/v_r. If the r_t are autocorrelated, then the effective variance is v_re v_r (1 + ϱ)/(1 - ϱ). The theory is applied to populations of checkerspot butterfly, grizzly bear, wolf, and mountain lion.     

Survival analysis of stochastic single-species population models in polluted environments

This paper reports on the behaviors of single species models with and without pollution. We consider three basic models, one is deterministic, and others are stochastic. We first obtained the acute thresholds between local extinction and (stochastic) weakly persistent in the mean for population respectively. Then we study the attainability of population size 0 for the stochastic cases and show that a randomized non-autonomous logistic equation will be stochastic permanent under some conditions. Finally, we introduce some numerical simulink graphics to illustrate our main results.     

Incorporating Uncertainty into Management Models for Marine Mammals

Abstract: Good management models and good models for understanding biology differ in basic philosophy. Management models must facilitate management decisions despite large amounts of uncertainty about the managed populations. Such models must be based on parameters that can be estimated readily, must explicitly account for uncertainty, and should be simple to understand and implement. In contrast, biological models are designed to elucidate the workings of biology and should not be constrained by management concerns. We illustrate the need to incorporate uncertainty in management models by reviewing the inadequacy of using standard biological models to manage marine mammals in the United States. Past management was based on a simple model that, although it may have represented population dynamics adequately, failed as a management tool because the parameter that triggered management action was extremely difficult to estimate for the majority of populations. Uncertainty in parameter estimation resulted in few conservation actions. We describe a recently adopted management scheme that incorporates uncertainty and its resulting implementation. The approach used in this simple management scheme, which was tested by using simulation models, incorporates uncertainty and mandates monitoring abundance and human-caused mortality. Although the entire scheme may be suitable for application to some terrestrial and marine problems, two features are broadly applicable: the incorporation of uncertainty through simulations of management and the use of quantitative management criteria to translate verbal objectives into levels of acceptable risk.     

Metapopulation Dynamics and Amphibian Conservation

Abstract: In many respects, amphibian spatial dynamics resemble classical metapopulation models, in which subpopulations in breeding ponds blink in and out of existence and extinction and colonization rates are functions of pond spatial arrangement. This "ponds-as-patches" view of amphibian spatial dynamics is useful in several respects. First, it highlights the importance of regional and landscape processes in determining local patterns of abundance. Second, it offers a straightforward, pond-based approach to monitoring and managing amphibian populations. For many species, however, the ponds-as-patches view may be an oversimplification and metapopulation structure may be more apparent than real. Changes in distribution may be caused by processes other than extinction and recolonization, and most extinctions probably result from deterministic factors, not stochastic processes. In addition, the effects of pond isolation appear to be important primarily in disturbed environments, and in many cases these isolation effects may be better explained by the distribution of terrestrial habitats than by the distribution of breeding ponds. These complications have important implications for both researchers and managers. For researchers, future efforts need to determine the mechanisms underlying patterns of abundance and distributional change and patterns in amphibian populations. For managers, effective conservation strategies must successfully balance metapopulation considerations with careful attention to local habitat quality. Furthermore, translocations and active management may be indispensable tools for conserving amphibians in landscapes containing multiple breeding ponds.     

Incorporating known sources of uncertainty to determine precautionary harvests of saltwater crocodiles.

It has been demonstrated repeatedly that the degree to which regulation operates and the magnitude of environmental variation in an exploited population will together dictate the type of sustainable harvest achievable. Yet typically, harvest models fail to incorporate uncertainty in the underlying dynamics of the target population by assuming a particular (unknown) form of endogenous control. We use a novel approach to estimate the sustainable yield of saltwater crocodile (Crocodylus porosus) populations from major river systems in the Northern Territory, Australia, as an example of a system with high uncertainty. We used multimodel inference to incorporate three levels of uncertainty in yield estimation: (1) uncertainty in the choice of the underlying model(s) used to describe population dynamics, (2) the error associated with the precision and bias of model parameter estimation, and (3) environmental fluctuation (process error). We demonstrate varying strength of evidence for density regulation (1.3-96.7%) for crocodiles among 19 river systems by applying a continuum of five dynamical models (density-independent with and without drift and three alternative density-dependent models) to time series of density estimates. Evidence for density dependence increased with the number of yearly transitions over which each river system was monitored. Deterministic proportional maximum sustainable yield (PMSY) models varied widely among river systems (0.042-0.611), and there was strong evidence for an increasing PMSY as support for density dependence rose. However, there was also a large discrepancy between PMSY values and those produced by the full stochastic simulation projection incorporating all forms of uncertainty, which can be explained by the contribution of process error to estimates of sustainable harvest. We also determined that a fixed-quota harvest strategy (up to 0.2K, where K is the carrying capacity) reduces population size much more rapidly than proportional harvest (the latter strategy requiring temporal monitoring of population size to adjust harvest quotas) and greatly inflates the risk of resource depletion. Using an iconic species recovering from recent extreme overexploitation to examine the potential for renewed sustainable harvest, we have demonstrated that incorporating major forms of uncertainty into a single quantitative framework provides a robust approach to modeling the dynamics of exploited populations.