Population responses to perturbations: the importance of trait-based analysis illustrated through a microcosm experiment

Is Evolution in Response to Extreme Events Good for Population Persistence?

AbstractClimate change is predicted to increase the severity of environmental perturbations, including storms and droughts, which act as strong selective agents. These extreme events are often of finite duration (pulse disturbances). Hence, while evolution during an extreme event may be adaptive, the resulting phenotypic changes may become maladaptive when the event ends. Using individual-based models and analytic approximations that fuse quantitative genetics and demography, we explore how heritability and phenotypic variance affect population size and extinction risk in finite populations under an extreme event of fixed duration. Since more evolution leads to greater maladaptation and slower population recovery following an extreme event, greater heritability can increase extinction risk when the extreme event is short. Alternatively, when an extreme event is sufficiently long, heritability often helps a population persist. We also find that when events are severe, the buffering effect of phenotypic variance can outweigh the increased load it causes.

Movements of marine and estuarine turtles during Hurricane Michael

Natural disturbances are an important driver of population dynamics. Because it is difficult to observe wildlife during these events, our understanding of the strategies that species use to survive these disturbances is limited. On October 10, 2018, Hurricane Michael made landfall on Florida’s northwest coast. Using satellite and acoustic telemetry, we documented movements of 6 individual turtles: one loggerhead sea turtle, one Kemp’s ridley sea turtle, three green sea turtles and one diamondback terrapin, in a coastal bay located less than 30 km from hurricane landfall. Post-storm survival was confirmed for all but the Kemp’s ridley; the final condition of that individual remains unknown. No obvious movements were observed for the remaining turtles however the loggerhead used a larger home range in the week after the storm. This study highlights the resiliency of turtles in response to extreme weather conditions. However, long-term impacts to these species from habitat changes post-hurricane are unknown.

Corridor quality affects net movement, size of dispersers, and population growth in experimental microcosms

Corridors are expected to increase species dispersal in fragmented habitats. However, it remains unclear how the quality of corridors influences the dispersal process, and how it interacts with corridor length and width. Here we investigate these factors using a small-scale laboratory system where we track the dispersal of the model organism Collembola Folsomia candida. Using this system, we study the effects of corridor length, width, and quality on the probability of dispersal, net movement, body size of dispersers, and the rate of change in population size after colonization. We show that corridor quality positively affected dispersal probability, net movement, and the rate of change in population size in colonised patches. Moreover, corridor quality significantly affected the size of dispersers, with only larger individuals dispersing through poor quality corridors. The length and width of corridors affected both the rate at which populations increased in colonised patches and the net number of individuals which dispersed, suggesting that these physical properties may be important in maintaining the flow of individuals in space. Our results thus suggest that corridor quality can have an important role in determining not only the probability of dispersal occurs but also the phenotypes of the individuals which disperse, with concomitant effects on the net movement of individuals and the rate of change in population size in the colonised patches.

Investigating the Dynamics of Elk Population Size and Body Mass in a Seasonal Environment Using a Mechanistic Integral Projection Model

Environmentally mediated changes in body size often underlie population responses to environmental change, yet this is not a universal phenomenon. Understanding when phenotypic change underlies population responses to environmental change is important for obtaining insights and robust predictions of population dynamics in a changing world. We develop a dynamic integral projection model that mechanistically links environmental conditions to demographic rates and phenotypic traits (body size) via changes in resource availability and individual energetics. We apply the model to the northern Yellowstone elk population and explore population responses to changing patterns of seasonality, incorporating the interdependence of growth, demography, and density-dependent processes operating through population feedback on available resources. We found that small changes in body size distributions can have large impacts on population dynamics but need not cause population responses to environmental change. Environmental changes that altered demographic rates directly, via increasing or decreasing resource availability, led to large population impacts in the absence of substantial changes to body size distributions. In contrast, environmentally driven shifts in body size distributions could occur with little consequence for population dynamics when the effect of environmental change on resource availability was small and seasonally restricted and when strong density-dependent processes counteracted expected population responses. These findings highlight that a robust understanding of how associations between body size and demography influence population responses to environmental change will require knowledge of the shape of the relationship between phenotypic distributions and vital rates, the population status with regard to its carrying capacity, and importantly the nature of the environmentally driven change in body size and carrying capacity.

The Evolution of Variance Control

Genetically identical individuals can be phenotypically variable, even in constant environmental conditions. The ubiquity of this phenomenon, known as 'intra-genotypic variability', is increasingly evident and the relevant mechanistic underpinnings are beginning to be understood. In parallel, theory has delineated a number of formal expectations for contexts in which such a feature would be adaptive. Here, we review empirical evidence across biological systems and theoretical expectations, including nonlinear averaging and bet hedging. We synthesize existing results to illustrate the dependence of selection outcomes both on trait characteristics, features of environmental variability, and species' demographic context. We conclude by discussing ways to bridge the gap between empirical evidence of intra-genotypic variability, studies demonstrating its genetic component, and evidence that it is adaptive.

Fullerene toxicity in the benthos with implications for freshwater ecosystem services

Production of engineered carbon-based nanomaterials (CNMs) is rising, with increased risk of release to the environment during production, use, and disposal. This trend highlights a need to understand potential impacts of CNMs on the natural environment. Fullerenes are an emerging class of CNMs that are insoluble in water, and form aggregates that settle quickly, suggesting higher relative vulnerability of aquatic benthic ecosystems. This study assessed eco-toxicity of fullerenes (C60, C70) and the functionalized derivative, phenyl-C61-butyric acid methyl ester (PCBM), on functionally representative benthic organisms in traditional laboratory assays, and evaluated how the potential lethal and sub-lethal effects of fullerenes may indirectly impact benthic ecosystem function, including decomposition, primary productivity and nutrient cycling in lake microcosms with natural sediments. Standard toxicity tests indicated that population growth of Lumbriculus variegatus was reduced at 25 to 150 mg C60 kg^-1, but C70 and PCBM did not affect growth or weight of organisms in artificial sediments at 25 mg kg^-1. Survivorship and growth were lower in natural sediments with historic contamination, but C60 did not exacerbate this effect. C60 inhibited photosynthesis by the benthic diatom Nitzschia palea, and at high exposure chlorophyll a increased, suggesting a shading response. L. variegatus had strong effects on benthic ecosystem function, especially metabolism and nitrogen cycling, but C60 ≤ 30 mg kg^-1 sediment did not influence the role of L. variegatus in driving benthic processes. These observations suggest that at moderate to high concentrations, C60 may directly impact benthic organisms. However, under natural conditions with low to moderate concentrations, C60 has little effect and does not indirectly impact the ecosystem processes maintained by such organisms. These results are a step further towards a better understanding of potential impacts of fullerenes on aquatic ecosystems, and can aid in the development of regulatory policies.

Trait-based predictions and responses from laboratory mite populations to harvesting in stochastic environments

Predictions on population responses to perturbations are often derived from trait-based approaches like integral projection models (IPMs), but are rarely tested. IPMs are constructed from functions that describe survival, growth and reproduction in relation to the traits of individuals and their environment. Although these functions comprise biologically non-informative statistical coefficients within standard IPMs, model parameters of the recently developed dynamic energy budget IPM (DEB-IPM) are life-history traits like "length at maturation" and "maximum reproduction rate". Testing predictions from mechanistic IPMs against empirical observations can therefore provide functional insights into the links between individual life history, the environment and population dynamics. Here, we compared the population dynamics of the bulb mite (Rhizoglyphus robini) predicted by a DEB-IPM with those observed in an experiment where populations experienced daily food rations that were either positively correlated over time (red noise), negatively (blue noise) or uncorrelated (white noise). We also selectively harvested large adults in half of these populations. The model failed to generate detailed predictions of population structure as juvenile numbers were overestimated; likely because juvenile-adult interference competition was underestimated. The model performed well at the population level as, for both harvested and unharvested populations, simulations matched the observed, long-term stochastic growth rate λ_s . We next generalised the model to investigate how stochastic change affects mite λ_s , which correlated well with the frequency f of experiencing periods of good environment, but, due to the relationship between f and noise colour ρ, did not correlate well with shifts in ρ. The sensitivity of λ_s to perturbations in life-history parameters depended on the type of stochastic change, as well as population growth. Our findings show that responses to differential mortality depend on individual life-history traits, environmental characteristics and population growth. As long-term climate change causes ever greater environmental fluctuations, trait-based approaches will be increasingly important in predicting population responses to change. We therefore conclude by illustrating what questions can be examined with mechanistic trait-based models like the DEB-IPM, the answers to which will advance our knowledge of the functional links between individual traits, the environment and population dynamics.

When Do Shifts in Trait Dynamics Precede Population Declines?

Predicting population responses to environmental change is an ongoing challenge in ecology. Studies investigating the links between fitness-related phenotypic traits and demography have shown that trait dynamic responses to environmental change can sometimes precede population dynamic responses and thus can be used as an early warning signal. However, it is still unknown under which ecological and evolutionary circumstances shifts in fitness-related traits can precede population responses to environmental perturbation. Here, we take a trait-based demographic approach and investigate both trait and population dynamics in a density-regulated population in response to a gradual change in the environment. We explore the ecological and evolutionary constraints under which shifts in fitness-related traits precede a decline in population size. We show both analytically and with experimental data that under medium to slow rates of environmental change, shifts in a trait value can precede population decline. We further show the positive influence of environmental predictability, net reproductive rate, plasticity, and genetic variation on shifts in trait dynamics preceding potential population declines. These results still hold under nonconstant genetic variation and environmental stochasticity. Our study highlights ecological and evolutionary circumstances under which a fitness-related trait can be used as an early warning signal of an impending population decline.

Demographic responses underlying eco-evolutionary dynamics as revealed with inverse modelling

Changes in population dynamics due to interacting evolutionary and ecological processes are the direct result of responses in vital rates, that is stage‐specific growth, survival and fecundity. Quantifying through which vital rates population fitness is affected, instead of focusing on population trends only, can give a more mechanistic understanding of eco‐evolutionary dynamics.

The aim of this study was to estimate the underlying demographic rates of aphid (Myzus persicae) populations. We analysed unpublished stage‐structure population dynamics data of a field experiment with caged and uncaged populations in which rapid evolutionary dynamics were observed, as well as unpublished results from an individual life table experiment performed in a glasshouse.

Using data on changes in population abundance and stage distributions over time, we estimated transition matrices with inverse modelling techniques, in a Bayesian framework. The model used to fit across all experimental treatments included density as well as clone‐specific caging effects. We additionally used individual life table data to inform the model on survival, growth and reproduction.

Results suggest that clones varied considerably in vital rates, and imply trade‐offs between reproduction and survival. Responses to densities also varied between clones. Negative density dependence was found in growth and reproduction, and the presence of predators and competitors further decreased these two vital rates, while survival estimates increased. Under uncaged conditions, population growth rates of the evolving populations were increased compared to the expectation based on the pure clones.

Our inverse modelling approach revealed how much vital rates contributed to the eco‐evolutionary dynamics. The decomposition analysis showed that variation in population growth rates in the evolving populations was to a large extent shaped by plant size. Yet, it also revealed an impact of evolutionary changes in clonal composition. Finally, we discuss that inverse modelling is a complex problem, as multiple combinations of individual rates can result in the same dynamics. We discuss assumptions and limitations, as well as opportunities, of this approach.

Stochastic Evolutionary Demography under a Fluctuating Optimum Phenotype

Many natural populations exhibit temporal fluctuations in abundance that are consistent with external forcing by a randomly changing environment. As fitness emerges from an interaction between the phenotype and the environment, such demographic fluctuations probably include a substantial contribution from fluctuating phenotypic selection. We study the stochastic population dynamics of a population exposed to random (plus possibly directional) changes in the optimum phenotype for a quantitative trait that evolves in response to this moving optimum. We derive simple analytical predictions for the distribution of log population size over time both transiently and at stationarity under Gompertz density regulation. These predictions are well matched by population- and individual-based simulations. The log population size is approximately reverse gamma distributed, with a negative skew causing an excess of low relative to high population sizes, thus increasing extinction risk relative to a symmetric (e.g., normal) distribution with the same mean and variance. Our analysis reveals how the mean and variance of log population size change with the variance and autocorrelation of deviations of the evolving mean phenotype from the optimum. We apply our results to the analysis of evolutionary rescue in a stochastic environment and show that random fluctuations in the optimum can substantially increase extinction risk by both reducing the expected growth rate and increasing the variance of population size by several orders of magnitude.

Harvested populations are more variable only in more variable environments

The interaction between environmental variation and population dynamics is of major importance, particularly for managed and economically important species, and especially given contemporary changes in climate variability. Recent analyses of exploited animal populations contested whether exploitation or environmental variation has the greatest influence on the stability of population dynamics, with consequences for variation in yield and extinction risk. Theoretical studies however have shown that harvesting can increase or decrease population variability depending on environmental variation, and requested controlled empirical studies to test predictions. Here, we use an invertebrate model species in experimental microcosms to explore the interaction between selective harvesting and environmental variation in food availability in affecting the variability of stage‐structured animal populations over 20 generations. In a constant food environment, harvesting adults had negligible impact on population variability or population size, but in the variable food environments, harvesting adults increased population variability and reduced its size. The impact of harvesting on population variability differed between proportional and threshold harvesting, between randomly and periodically varying environments, and at different points of the time series. Our study suggests that predicting the responses to selective harvesting is sensitive to the demographic structures and processes that emerge in environments with different patterns of environmental variation.

Including trait-based early warning signals helps predict population collapse

Foreseeing population collapse is an on-going target in ecology, and this has led to the development of early warning signals based on expected changes in leading indicators before a bifurcation. Such signals have been sought for in abundance time-series data on a population of interest, with varying degrees of success. Here we move beyond these established methods by including parallel time-series data of abundance and fitness-related trait dynamics. Using data from a microcosm experiment, we show that including information on the dynamics of phenotypic traits such as body size into composite early warning indices can produce more accurate inferences of whether a population is approaching a critical transition than using abundance time-series alone. By including fitness-related trait information alongside traditional abundance-based early warning signals in a single metric of risk, our generalizable approach provides a powerful new way to assess what populations may be on the verge of collapse.

Predicting population collapse by monitoring key early warning signals in time-series data may highlight when interventions are needed. Here, the authors show that including information on phenotypic traits like body size can more accurately predict critical transitions than abundance data alone.

How well can body size represent effects of the environment on demographic rates? Disentangling correlated explanatory variables

Demographic rates are shaped by the interaction of past and current environments that individuals in a population experience. Past environments shape individual states via selection and plasticity, and fitness-related traits (e.g. individual size) are commonly used in demographic analyses to represent the effect of past environments on demographic rates. We quantified how well the size of individuals captures the effects of a population's past and current environments on demographic rates in a well-studied experimental system of soil mites. We decomposed these interrelated sources of variation with a novel method of multiple regression that is useful for understanding nonlinear relationships between responses and multicollinear explanatory variables. We graphically present the results using area-proportional Venn diagrams. Our novel method was developed by combining existing methods and expanding upon them. We showed that the strength of size as a proxy for the past environment varied widely among vital rates. For instance, in this organism with an income breeding life history, the environment had more effect on reproduction than individual size, but with substantial overlap indicating that size encompassed some of the effects of the past environment on fecundity. This demonstrates that the strength of size as a proxy for the past environment can vary widely among life-history processes within a species, and this variation should be taken into consideration in trait-based demographic or individual-based approaches that focus on phenotypic traits as state variables. Furthermore, the strength of a proxy will depend on what state variable(s) and what demographic rate is being examined; that is, different measures of body size (e.g. length, volume, mass, fat stores) will be better or worse proxies for various life-history processes.

Building integral projection models: a user's guide

In order to understand how changes in individual performance (growth, survival or reproduction) influence population dynamics and evolution, ecologists are increasingly using parameterized mathematical models. For continuously structured populations, where some continuous measure of individual state influences growth, survival or reproduction, integral projection models (IPMs) are commonly used. We provide a detailed description of the steps involved in constructing an IPM, explaining how to: (i) translate your study system into an IPM; (ii) implement your IPM; and (iii) diagnose potential problems with your IPM. We emphasize how the study organism's life cycle, and the timing of censuses, together determine the structure of the IPM kernel and important aspects of the statistical analysis used to parameterize an IPM using data on marked individuals. An IPM based on population studies of Soay sheep is used to illustrate the complete process of constructing, implementing and evaluating an IPM fitted to sample data. We then look at very general approaches to parameterizing an IPM, using a wide range of statistical techniques (e.g. maximum likelihood methods, generalized additive models, nonparametric kernel density estimators). Methods for selecting models for parameterizing IPMs are briefly discussed. We conclude with key recommendations and a brief overview of applications that extend the basic model. The online Supporting Information provides commented R code for all our analyses.

Correlative changes in life-history variables in response to environmental change in a model organism

Global change alters the environment, including increases in the frequency of (un)favorable events and shifts in environmental noise color. However, how these changes impact the dynamics of populations, and whether these can be predicted accurately has been largely unexamined. Here we combine recently developed population modeling approaches and theory in stochastic demography to explore how life history, morphology, and average fitness respond to changes in the frequency of favorable environmental conditions and in the color of environmental noise in a model organism (an acarid mite). We predict that different life-history variables respond correlatively to changes in the environment, and we identify different life-history variables, including lifetime reproductive success, as indicators of average fitness and life-history speed across stochastic environments. Depending on the shape of adult survival rate, generation time can be used as an indicator of the response of populations to stochastic change, as in the deterministic case. This work is a useful step toward understanding population dynamics in stochastic environments, including how stochastic change may shape the evolution of life histories.

Eco-evolutionary dynamics in response to selection on life-history

Understanding the consequences of environmental change on ecological and evolutionary dynamics is inherently problematic because of the complex interplay between them. Using invertebrates in microcosms, we characterise phenotypic, population and evolutionary dynamics before, during and after exposure to a novel environment and harvesting over 20 generations. We demonstrate an evolved change in life-history traits (the age- and size-at-maturity, and survival to maturity) in response to selection caused by environmental change (wild to laboratory) and to harvesting (juvenile or adult). Life-history evolution, which drives changes in population growth rate and thus population dynamics, includes an increase in age-to-maturity of 76% (from 12.5 to 22 days) in the unharvested populations as they adapt to the new environment. Evolutionary responses to harvesting are outweighed by the response to environmental change (∼ 1.4 vs. 4% change in age-at-maturity per generation). The adaptive response to environmental change converts a negative population growth trajectory into a positive one: an example of evolutionary rescue.