Elasticity analysis of density-dependent matrix population models: the invasion exponent and its substitutes

In density-independent models, the population growth rate lambda measures population performance, and the perturbation analysis of lambda (its sensitivity and elasticity) plays an important role in demography. In density-dependent models, the invasion exponent lambdaI replaces lambda as a measure of population performance. The perturbation analysis of lambdaI reveals the effects of environmental changes and management actions, gives the direction and intensity of density-dependent natural selection on life history traits, and permits calculation of the sampling variance of the invasion exponent. Because density-dependent models require more data than density-independent models, it is tempting to look for substitutes for the invasion exponent, the sensitivity and elasticity of which can be calculated from a density-independent model. Here we examine the accuracy of two such substitutes: the dominant eigenvalue of the projection matrix evaluated at equilibrium (An) and the dominant eigenvalue of the matrix averaged over the attractor (A). Using a two-stage model that represents a wide range of life history types, we find that the elasticities of An or A often agree to within less than 5% error with those of the invasion exponent, even when population dynamics are chaotic. The exceptions are for semelparous life histories, especially when density-dependence affects fertility. This suggests that the elasticity analysis of density-independent models near equilibrium, or averaged over the attractor, provides useful information about the elasticity of the invasion exponent in density-dependent models.     

Individual heterogeneity in life histories and eco‐evolutionary dynamics

Individual heterogeneity in life history shapes eco‐evolutionary processes, and unobserved heterogeneity can affect demographic outputs characterising life history and population dynamical properties. Demographic frameworks like matrix models or integral projection models represent powerful approaches to disentangle mechanisms linking individual life histories and population‐level processes. Recent developments have provided important steps towards their application to study eco‐evolutionary dynamics, but so far individual heterogeneity has largely been ignored. Here, we present a general demographic framework that incorporates individual heterogeneity in a flexible way, by separating static and dynamic traits (discrete or continuous). First, we apply the framework to derive the consequences of ignoring heterogeneity for a range of widely used demographic outputs. A general conclusion is that besides the long‐term growth rate lambda, all parameters can be affected. Second, we discuss how the framework can help advance current demographic models of eco‐evolutionary dynamics, by incorporating individual heterogeneity. For both applications numerical examples are provided, including an empirical example for pike. For instance, we demonstrate that predicted demographic responses to climate warming can be reversed by increased heritability. We discuss how applications of this demographic framework incorporating individual heterogeneity can help answer key biological questions that require a detailed understanding of eco‐evolutionary dynamics.     

Modeling vital rates improves estimation of population projection matrices

Population projection matrices are commonly used by ecologists and managers to analyze the dynamics of stage-structured populations. Building projection matrices from data requires estimating transition rates among stages, a task that often entails estimating many parameters with few data. Consequently, large sampling variability in the estimated transition rates increases the uncertainty in the estimated matrix and quantities derived from it, such as the population multiplication rate and sensitivities of matrix elements. Here, we propose a strategy to avoid overparameterized matrix models. This strategy involves fitting models to the vital rates that determine matrix elements, evaluating both these models and ones that estimate matrix elements individually with model selection via information criteria, and averaging competing models with multimodel averaging. We illustrate this idea with data from a population of Silene acaulis (Caryophyllaceae), and conduct a simulation to investigate the statistical properties of the matrices estimated in this way. The simulation shows that compared with estimating matrix elements individually, building population projection matrices by fitting and averaging models of vital-rate estimates can reduce the statistical error in the population projection matrix and quantities derived from it.     

Partial life cycle analysis: a model for pre-breeding census data

Matrix population models have become popular tools in research areas as diverse as population dynamics, life history theory, wildlife management, and conservation biology. Two classes of matrix models are commonly used for demographic analysis of age‐structured populations: age‐structured (Leslie) matrix models, which require age‐specific demographic data, and partial life cycle models, which can be parameterized with partial demographic data. Partial life cycle models are easier to parameterize because data needed to estimate parameters for these models are collected much more easily than those needed to estimate age‐specific demographic parameters. Partial life cycle models also allow evaluation of the sensitivity of population growth rate to changes in ages at first and last reproduction, which cannot be done with age‐structured models. Timing of censuses relative to the birth‐pulse is an important consideration in discrete‐time population models but most existing partial life cycle models do not address this issue, nor do they allow fractional values of variables such as ages at first and last reproduction. Here, we fully develop a partial life cycle model appropriate for situations in which demographic data are collected immediately before the birth‐pulse (pre‐breeding census). Our pre‐breeding census partial life cycle model can be fully parameterized with five variables (age at maturity, age at last reproduction, juvenile survival rate, adult survival rate, and fertility), and it has some important applications even when age‐specific demographic data are available (e.g., perturbation analysis involving ages at first and last reproduction). We have extended the model to allow non‐integer values of ages at first and last reproduction, derived formulae for sensitivity analyses, and presented methods for estimating parameters for our pre‐breeding census partial life cycle model. We applied the age‐structured Leslie matrix model and our pre‐breeding census partial life cycle model to demographic data for several species of mammals. Our results suggest that dynamical properties of the age‐structured model are generally retained in our partial life cycle model, and that our pre‐breeding census partial life cycle model is an excellent proxy for the age‐structured Leslie matrix model.     

Effects of reproductive mode on demography and life history in Arabis fecunda (Brassicaceae)

Life history theory predicts that trade-offs between growth and reproduction should be dictated by a population's mortality schedule. We tested this prediction with Arabis fecunda, a short-lived perennial that occurs in many different habitats in southwest Montana. Individuals produce either or both axillary or terminal inflorescences. Axillary-flowering plants are usually iteroparous and have smaller reproductive bouts, while terminal-flowering plants have larger reproductive bouts, and tend to be semelparous. We recorded size and fecundity of A. fecunda individuals from 1989 to 1993 in three different habitats. There was great variation in demographic and life history traits among the populations. A wide range of life history strategies among populations of A. fecunda is achieved through different proportions of axillary- and terminal-flowering plants. Arabis fecunda demonstrated a lower recruitment rate, higher survivorship, slower growth, and lower annual fecundity at the low-elevation site compared to the high-elevation site. At the low-elevation site population size was more stable, and elasticity analysis of matrix projection models indicated that adult survivorship was the most important demographic parameter contributing to population growth. This association of life history characters conforms to theoretical predictions.     

What life cycle graphs can tell about the evolution of life histories

We analyze long-term evolutionary dynamics in a large class of life history models. The model family is characterized by discrete-time population dynamics and a finite number of individual states such that the life cycle can be described in terms of a population projection matrix. We allow an arbitrary number of demographic parameters to be subject to density-dependent population regulation and two or more demographic parameters to be subject to evolutionary change. Our aim is to identify structural features of life cycles and modes of population regulation that correspond to specific evolutionary dynamics. Our derivations are based on a fitness proxy that is an algebraically simple function of loops within the life cycle. This allows us to phrase the results in terms of properties of such loops which are readily interpreted biologically. The following results could be obtained. First, we give sufficient conditions for the existence of optimisation principles in models with an arbitrary number of evolving traits. These models are then classified with respect to their appropriate optimisation principle. Second, under the assumption of just two evolving traits we identify structural features of the life cycle that determine whether equilibria of the monomorphic adaptive dynamics (evolutionarily singular points) correspond to fitness minima or maxima. Third, for one class of frequency-dependent models, where optimisation is not possible, we present sufficient conditions that allow classifying singular points in terms of the curvature of the trade-off curve. Throughout the article we illustrate the utility of our framework with a variety of examples.     

Graph decompositions for demographic loop analysis

A new approach to loop analysis is presented in which decompositions of the total elasticity of a population projection matrix over a set of life history pathways are obtained as solutions of a constrained system of linear equations. In loop analysis, life history pathways are represented by loops in the life cycle graph, and the elasticity of the loop is interpreted as a measure of the contribution of the life history pathway to the population growth rate. Associated with the life cycle graph is a vector space -- the cycle space of the graph -- which is spanned by the loops. The elasticities of the transitions in the life cycle graph can be represented by a vector in the cycle space, and a loop decomposition of the life cycle graph is then defined to be any nonnegative linear combination of the loops which sum to the vector of elasticities. In contrast to previously published algorithms for carrying out loop analysis, we show that a given life cycle graph admits of either a unique loop decomposition or an infinite set of loop decompositions which can be characterized as a bounded convex set of nonnegative vectors. Using this approach, loop decompositions which minimize or maximize a linear objective function can be obtained as solutions of a linear programming problem, allowing us to place lower and upper bounds on the contributions of life history pathways to the population growth rate. Another consequence of our approach to loop analysis is that it allows us to identify the exact tradeoffs in contributions to the population growth rate that must exist between life history pathways.     

Demographic analysis of continuous-time life-history models

I present a computational approach to calculate the population growth rate, its sensitivity to life-history parameters and associated statistics like the stable population distribution and the reproductive value for exponentially growing populations, in which individual life history is described as a continuous development through time. The method is generally applicable to analyse population growth and performance for a wide range of individual life-history models, including cases in which the population consists of different types of individuals or in which the environment is fluctuating periodically. It complements comparable methods developed for discrete-time dynamics modelled with matrix or integral projection models. The basic idea behind the method is to use Lotka's integral equation for the population growth rate and compute the integral occurring in that equation by integrating an ordinary differential equation, analogous to recently derived methods to compute steady-states of physiologically structured population models. I illustrate application of the method using a number of published life-history models.     

Stochastic stable population growth in integral projection models: theory and application

Stochastic matrix projection models are widely used to model age- or stage-structured populations with vital rates that fluctuate randomly over time. Practical applications of these models rest on qualitative properties such as the existence of a long term population growth rate, asymptotic log-normality of total population size, and weak ergodicity of population structure. We show here that these properties are shared by a general stochastic integral projection model, by using results in (Eveson in D. Phil. Thesis, University of Sussex, 1991, Eveson in Proc. Lond. Math. Soc. 70, 411-440, 1993) to extend the approach in (Lange and Holmes in J. Appl. Prob. 18, 325-344, 1981). Integral projection models allow individuals to be cross-classified by multiple attributes, either discrete or continuous, and allow the classification to change during the life cycle. These features are present in plant populations with size and age as important predictors of individual fate, populations with a persistent bank of dormant seeds or eggs, and animal species with complex life cycles. We also present a case-study based on a 6-year field study of the Illyrian thistle, Onopordum illyricum, to demonstrate how easily a stochastic integral model can be parameterized from field data and then applied using familiar matrix software and methods. Thistle demography is affected by multiple traits (size, age and a latent "quality" variable), which would be difficult to accommodate in a classical matrix model. We use the model to explore the evolution of size- and age-dependent flowering using an evolutionarily stable strategy (ESS) approach. We find close agreement between the observed flowering behavior and the predicted ESS from the stochastic model, whereas the ESS predicted from a deterministic version of the model is very different from observed flowering behavior. These results strongly suggest that the flowering strategy in O. illyricum is an adaptation to random between-year variation in vital rates.     

Two paradigms of mathematical population biology. Attempting to synthesise

Whatever popular the slogan of nonlinear ecological interactions has been in theory, practical ecology professes the projection matrix paradigm, which is essentially linear, i.e., the linear matrix model paradigm for discrete-structured population dynamics. The dominant eigenvalue lamda1, of the projection matrix L is considered as a growth potential of the population. It provides for a quantitative measure of the fitness at which the species is adapted to the given environment, the measure being adequate and accurate, given the data of"identified individuals" type. The case of"identified individuals with unknown parents" bears uncertainty in the status-specific reproduction rates, which eliminates in a unique way (for a broad class of structures and life cycle graphs) by maximizing lamda1(L) under the constraints ensuing from the data and knowledge of species biology. The paradigm of linearity gives way to nonlinear models when modeled are species interactions, such as competition for shared resources, and where the outcome of interaction depends on the population structure of the competitors. This circumstance dictates a need for the synthesis of two paradigms, which is achieved in nonlinear matrix operators as models of interaction between the species whose populations are discrete-structured.     

Integral projection models perform better for small demographic data sets than matrix population models: a case study of two perennial herbs

1. Matrix population models are widely used to describe population dynamics, conduct population viability analyses and derive management recommendations for plant populations. For endangered or invasive species, management decisions are often based on small demographic data sets. Hence, there is a need for population models which accurately assess population performance from such small data sets.
2. We used demographic data on two perennial herbs with different life histories to compare the accuracy and precision of the traditional matrix population model and the recently developed integral projection model (IPM) in relation to the amount of data.
3. For large data sets both matrix models and IPMs produced identical estimates of population growth rate (λ). However, for small data sets containing fewer than 300 individuals, IPMs often produced smaller bias and variance for λ than matrix models despite different matrix structures and sampling techniques used to construct the matrix population models.
4. Synthesis and applications . Our results suggest that the smaller bias and variance of λ estimates make IPMs preferable to matrix population models for small demographic data sets with a few hundred individuals. These results are likely to be applicable to a wide range of herbaceous, perennial plant species where demographic fate can be modelled as a function of a continuous state variable such as size. We recommend the use of IPMs to assess population performance and management strategies particularly for endangered or invasive perennial herbs where little demographic data are available.     

Modelling life history strategies with capture–recapture data: Evolutionary demography of the water skink Eulamprus tympanum

Matrix population models, elasticity analysis and loop analysis can potentially provide powerful techniques for the analysis of life histories. Data from a capture–recapture study on a population of southern highland water skinks (Eulamprus tympanum) were used to construct a matrix population model. Errors in elasticities were calculated by using the parametric bootstrap technique. Elasticity and loop analyses were then conducted to identify the life history stages most important to fitness. The same techniques were used to investigate the relative importance of fast versus slow growth, and rapid versus delayed reproduction. Mature water skinks were long‐lived, but there was high immature mortality. The most sensitive life history stage was the subadult stage. It is suggested that life history evolution in E. tympanum may be strongly affected by predation, particularly by birds. Because our population declined over the study, slow growth and delayed reproduction were the optimal life history strategies over this period. Although the techniques of evolutionary demography provide a powerful approach for the analysis of life histories, there are formidable logistical obstacles in gathering enough high‐quality data for robust estimates of the critical parameters.     

Sensitivity of the population growth rate to demographic variability within and between phases of the disturbance cycle

For species in disturbance-prone ecosystems, vital rates (survival, growth and reproduction) often vary both between and within phases of the cycle of disturbance and recovery; some of this variation is imposed by the environment, but some may represent adaptation of the life history to disturbance. Anthropogenic changes may amplify or impede these patterns of variation, and may have positive or negative effects on population growth. Using stochastic population projection matrix models, we develop stochastic elasticities (proportional derivatives of the long-run population growth rate) to gauge the population effects of three types of change in demographic variability (changes in within- and between-disturbance-phase variability and phase-specific changes). Computing these elasticities for five species of disturbance-influenced perennial plants, we pinpoint demographic rates that may reveal adaptation to disturbance, and we demonstrate that species may differ in their responses to different types of changes in demographic variability driven by climate change.     

Analysis of phenotypic change in relation to climatic drivers in a population of Soay sheep Ovis aries

Climatic change has frequently been identified as a key driver of change in biological communities. These changes can take the form of alterations to population dynamics, phenotypic characters, genetics and the life history of organisms and can have impacts on entire ecosystems. This study presents a novel investigation of how changes in a large scale climatic index, the North Atlantic Oscillation (NAO) can influence population dynamics and phenotypic characters in a population of ungulates. We use an integral projection model combined with actual climate change predictions to project future body size distributions for a population of Soay sheep Ovis aries. The climate change predictions used to direct our model projections were taken from published results of climate models, covering a range of different emissions scenarios. Our model results showed that for positive changes in the mean NAO large population declines occurred simultaneously with increases in mean body weight. The exact direction and magnitude of changes to population dynamics and character distributions were dependent on the greenhouse gas emissions scenario and model used to predict the NAO. This study has demonstrated how integral projection models can use outputs of climate models to direct projections of population dynamics and phenotypic character distributions. This approach allows the results of this study to be placed within current climate change research. The nature of integral projection models means that this methodology can be easily applied to other populations. The model can also be easily updated when new climate change predictions become available, making it a useful tool for understanding potential population level responses to climatic change. Synthesis Understanding how changes in climate affect biological communities is a key component in predicting the future form of populations. Utilising a novel approach that incorporates climatic drivers (in this instance the winter North Atlantic Oscillation) into an integral projection model framework, we predict future Soay sheep dynamics under specific climate change scenarios. Tracking quantitative trait distributions and life history metrics, our results predict declining population size and increasing body weight for an increasingly positive winter North Atlantic Oscillation index, as predicted by climate models. This has important implications for future wildlife management strategies and linking demographic responses to climate change.     

Using Nested Models and Laboratory Data for Predicting Population Effects of Contaminants on Fish: A Step Toward a Bottom-Up Approach for Establishing Causality in Field Studies

Predicting the effects of contaminants on fish populations is difficult due to their complex life history and high interannual variation in their population abundances. We present an approach that extrapolates laboratory data on contaminant effects, including behavioral effects, to the population level by using a series of nested statistical and simulation models. The approach is illustrated using PCB effects on Atlantic croaker. Laboratory experiments were performed that estimated PCB effects on fecundity, egg mortality, and the swimming speed and predator evasion behavior of larvae. A statistical model converted impaired predator evasion to reduced probability of escaping a predatory fish. An individual-based model then converted the output of the statistical model into changes in larval stage duration and survival, which were used to change elements of the matrix model. A matrix projection model simulated population dynamics for 100 years for baseline conditions and for two hypothetical PCB exposure scenarios. PCB effects were imposed in the model by reducing the fecundity of exposed adults, increasing egg mortality, and increasing the larval stage duration and mortality rate. Predicted population effects of PCBs were small relative to the interannual variation. Our analysis is a step toward understanding population responses to stressors and for ultimately establishing causality in field situations.     

Matrix analysis to model the invasion of alligatorweed (Alternanthera philoxeroides) on Kentucky lakes

Aims Alligatorweed (Alternanthera philoxeroides (Mart.) Griseb.) is an invasive species indigenous to South America. With its rapid invasion of southeastern US waterways, understanding the invasiveness of this plant species is critical for providing possible mechanisms of prevention for resource managers. The aim of this project is to use a matrix model to study the invasion dynamics of alligatorweed under both terrestrial and aquatic environments. The use of this model allows for a deeper understanding of the invasiveness and life history–stage structure of alligatorweed. In particular, matrix analysis can further test the hypothesis that certain life stages of alligatorweed might be more sensitive to control and management.Methods A greenhouse experiment was conducted to study the spread of alligatorweed under both aquatic and terrestrial environments. Utilizing the growth data obtained during the summer of 2010, matrix analysis was used to model the growth of alligatorweed for six different treatments. Transition matrices were generated based on plant measurements taken at different life stages defined by the number of leaves or nodes. These matrices are population projection models whose eigenvalues represent the growth rate of alligatorweed. A high growth rate is a key feature of successful invaders. Residuals were calculated and sensitivity analysis was performed to test the accuracy of the model and importance of each life stage over the entire life cycle of alligatorweed.Important findings The results of this study indicate that in the aquatic habitat, plants at their early life cycle stage are most sensitive to potential control measures. Conversely, in the terrestrial habitat, the most sensitive stage of alligatorweed is at its late life cycle stage, characterized with large-sized plants, thus suggesting the best timing for management and eradication of this invasive species.     

Demography_Lab,an educational application to evaluate population growth: Unstructured and matrix models

Training in Population Ecology asks for scalable applications capable of embarking students on a trip from basic concepts to the projection of populations under the various effects of density dependence and stochasticity. Demography_Lab is an educational tool for teaching Population Ecology aspiring to cover such a wide range of objectives. The application uses stochastic models to evaluate the future of populations. Demography_Lab may accommodate a wide range of life cycles and can construct models for populations with and without an age or stage structure. Difference equations are used for unstructured populations and matrix models for structured populations. Both types of models operate in discrete time. Models can be very simple, constructed with very limited demographic information or parameter‐rich, with a complex density‐dependence structure and detailed effects of the different sources of stochasticity. Demography_Lab allows for deterministic projections, asymptotic analysis, the extraction of confidence intervals for demographic parameters, and stochastic projections. Stochastic population growth is evaluated using up to three sources of stochasticity: environmental and demographic stochasticity and sampling error in obtaining the projection matrix. The user has full control on the effect of stochasticity on vital rates. The effect of the three sources of stochasticity may be evaluated independently for each vital rate. The user has also full control on density dependence. It may be included as a ceiling population size controlling the number of individuals in the population or it may be evaluated independently for each vital rate. Sensitivity analysis can be done for the asymptotic population growth rate or for the probability of extinction. Elasticity of the probability of extinction may be evaluated in response to changes in vital rates, and in response to changes in the intensity of density dependence and environmental stochasticity.     

Sex in an uncertain world: environmental stochasticity helps restore competitive balance between sexually and asexually reproducing populations

Like many organisms, individuals of the freshwater ostracod species Eucypris virens exhibit either obligate sexual or asexual reproductive modes. Both types of individual routinely co‐occur, including in the same temporary freshwater pond (their natural habitat in which they undergo seasonal diapause). Given the well‐known two‐fold cost of sex, this begs the question of how sexually reproducing individuals are able to coexist with their asexual counterparts in spite of such overwhelming costs. Environmental stochasticity in the form of ‘false dawn’ inundations (where the first hydration is ephemeral and causes loss of early hatching individuals) may provide an advantage to the sexual subpopulation, which shows greater variation in hatching times following inundation. We explore the potential role of environmental stochasticity in this system using life‐history data analysis, climate data, and matrix projection models. In the absence of environmental stochasticity, the population growth rate is significantly lower in sexual subpopulations. Climate data reveal that ‘false dawn’ inundations are common. Using matrix projection modelling with and without environmental stochasticity, we demonstrate that this phenomenon can restore appreciable balance to the system, in terms of population growth rates. This provides support for the role of environmental stochasticity in helping to explain the maintenance of sex and the occurrence of geographical parthenogenesis.     

The usefulness of sensitivity analysis for predicting the effects of cat predation on the population dynamics of their avian prey

Sensitivity analyses of population projection matrix (PPM) models are often used to identify life-history perturbations that will most influence a population's future dynamics. Sensitivities are linear extrapolations of the relationship between a population's growth rate and perturbations to its demographic parameters. Their effectiveness depends on the validity of the assumption of linearity. Here we assess whether sensitivity analysis is an appropriate tool to investigate the effect of predation by cats on the population growth rates of their avian prey. We assess whether predation by cats leads to non-linear effects on population growth and compare population growth rates predicted by sensitivity analysis with those predicted by a non-linear simulation. For a two-stage, age-classified House Sparrow Passer domesticus PPM slight non-linearity arose when PPM elements were perturbed, but perturbation to the vital rates underlying the matrix elements had a linear impact on population growth rate. We found a similar effect with a slightly larger three-stage, age-classified PPM for a Winter Wren Troglodytes troglodytes population perturbed by cat predation. For some avian species, predation by cats may cause linear or only slightly nonlinear impacts on population growth rates. For these species, sensitivity analysis appears to be a useful conservation tool. However, further work on multiple perturbations to avian prey species with more complicated life histories and higher-dimension PPM models is required.