Metapopulation persistence in heterogeneous landscapes: lessons about the effect of stochasticity

This article addresses an important aspect of the analysis of metapopulation persistence. It highlights some consequences of ignoring and including stochasticity in the sequence of extinction and colonization events. The results are based on a comparative analysis of the outcomes of two (one deterministic, one stochastic) spatially realistic metapopulation models and a search for common effects and differences. One key result of the article is that, under certain conditions, there are extra effects of the landscape structure (number and configuration of patches, patch size distribution) on metapopulation persistence if stochasticity is included. In these cases, ignoring or including stochasticity can change conclusions about the persistence status but also ranking orders, relative results, and qualitative trends. A list of conditions is provided under which including stochasticity is vital to prevent counterproductive conclusions about metapopulation persistence. The results of the overall study are condensed in five lessons about the effect of stochasticity. A number of implications for ecological theory and conservation management are discussed. The study demonstrates the potential of three recently published approximation formulas (metapopulation capacity lambdaM, mean lifetime Tm, and effective number of patches N) to serve as tools for ecological analysis and thinking.     

Metapopulation persistence in fragmented landscapes: significant interactions between genetic and demographic processes

We formulated a mathematical model in order to study the joint influence of demographic and genetic processes on metapopulation viability. Moreover, we explored the influence of habitat structure, matrix quality and disturbance on the interplay of these processes. We showed that the conditions that allow metapopulation persistence under the synergistic action of genetic and demographic processes depart significantly from predictions based on a mere superposition of the effects of each process separately. Moreover, an optimal dispersal rate exists that maximizes the range of survival rates of dispersers under which metapopulation persists and at the same time allows the largest sustainable patch removal and patch‐size reduction. The relative impact of patch removal and patch‐size reduction depends both on matrix quality and the dispersal strategy of the species: metapopulation persistence is more affected by patch‐size reduction (patch removal) for low (high)‐dispersing species, in presence of a low (high) quality matrix. Avoidance of inbreeding, through increased dispersal when the rate of inbreeding in a population is large, has positive effects on low‐dispersing species, but impairs the persistence of high‐dispersing species. Finally, size heterogeneity between patches largely influences metapopulation dynamics; the presence of large patches, even at the expense of other patches being smaller, can have positive effects on persistence in particular for species of low dispersing ability.     

Evolution of dispersal polymorphism and local adaptation of dispersal distance in spatially structured landscapes

Many organisms show polymorphism in dispersal distance strategies. This variation is particularly ecological relevant if it encompasses a functional separation of short‐ (SDD) and long‐distance dispersal (LDD). It remains, however, an open question whether both parts of the dispersal kernel are similarly affected by landscape related selection pressures. We implemented an individual‐based model to analyze the evolution of dispersal traits in fractal landscapes that vary in the proportion of habitat and its spatial configuration. Individuals are parthenogenetic with dispersal distance determined by two alleles on each individual's genome: one allele coding for the probability of global dispersal and one allele coding for the variance σ of a Gaussian local dispersal with mean value zero. Simulations show that mean distances of local dispersal and the probability of global dispersal, increase with increasing habitat availability, but that changes in the habitat's spatial autocorrelation impose opposing selective pressure: local dispersal distances decrease and global dispersal probabilities increase with decreasing spatial autocorrelation of the available habitat. Local adaptation of local dispersal distance emerges in landscapes with less than 70% of clumped habitat. These results demonstrate that long and short distance dispersal evolve separately according to different properties of the landscape. The landscape structure may consequently largely affect the evolution of dispersal distance strategies and the level of dispersal polymorphism.     

Metapopulation moments: coupling, stochasticity and persistence

1.  Spatial heterogeneity has long been viewed as a reliable means of increasing persistence. Here, an analytical model is developed to consider the variation and, hence, the persistence of stochastic metapopulations. This model relies on a novel moment closure technique, which is equivalent to assuming log-normal distributions for the population sizes.
2.  Single-species models show the greatest persistence when the mixing between subpopulations is large, so spatial heterogeneity is of no benefit. This result is confirmed by stochastic simulation of the full metapopulation.
3.  In contrast, natural-enemy models exhibit the greatest persistence for intermediate levels of coupling. When the coupling is too low, there are insufficient rescue effects between the subpopulations to sustain the dynamics, whereas when the coupling is too high all spatial heterogeneity is lost.
4.  The difference in behaviour between the one- and two-species models can be attributed to the oscillatory nature of the natural-enemy system.     

The role of landscape-dependent disturbance and dispersal in metapopulation persistence

The fundamental processes that influence metapopulation dynamics (extinction and recolonization) will often depend on landscape structure. Disturbances that increase patch extinction rates will frequently be landscape dependent such that they are spatially aggregated and have an increased likelihood of occurring in some areas. Similarly, landscape structure can influence organism movement, producing asymmetric dispersal between patches. Using a stochastic, spatially explicit model, we examine how landscape-dependent correlations between dispersal and disturbance rates influence metapopulation dynamics. Habitat patches that are situated in areas where the likelihood of disturbance is low will experience lower extinction rates and will function as partial refuges. We discovered that the presence of partial refuges increases metapopulation viability and that the value of partial refuges was contingent on whether dispersal was also landscape dependent. Somewhat counterintuitively, metapopulation viability was reduced when individuals had a preponderance to disperse away from refuges and was highest when there was biased dispersal toward refuges. Our work demonstrates that landscape structure needs to be incorporated into metapopulation models when there is either empirical data or ecological rationale for extinction and/or dispersal rates being landscape dependent.     

Impact of dispersal on population growth: the role of inter-patch distance

Dispersal can play a key role in the dynamics of patchy populations through patch colonization, and generally this leads to distance-dependent colonization. Less recognised are the roles of dispersal and inter-patch distance on the growth of a population after colonization. We use a laboratory mite model system in which both juveniles and adults can disperse to explore the impact of dispersal, and particularly inter-patch distance, on population dynamics. We examine the dynamics of patches after colonization by manipulating the presence of a dispersal corridor to a source patch at two inter-patch distances. Consistent with many field studies, the results show colonization was slower in more distant patches. Following colonization, the effect of the dispersal corridor on dynamics was dependent on inter-patch distance. In patches near the source, the number of adults tended to increase at a faster rate, and juveniles at a slower rate when connected with a dispersal corridor. In contrast, adult numbers grew slower and juveniles tended to grow faster when connected with a corridor in more distant patches. In the long-term, equilibrium adult numbers were lower in patches connected to the source patch at both distances. These results are likely to be driven by the effects of inter-patch distance on dispersal mortality, and the effects of dispersal on patch abundance and within-patch competition. These results confirm that distance is important for patch colonization and also show that distance can affect population density after colonization. The effects of dispersal and distance on local dynamics could be important in the dynamics of patchy populations in increasingly fragmented landscapes.     

Migration load in plants: role of pollen and seed dispersal in heterogeneous landscapes

Evolution of local adaptation depends critically on the level of gene flow, which, in plants, can be due to either pollen or seed dispersal. Using analytical predictions and individual-centred simulations, we investigate the specific influence of seed and pollen dispersal on local adaptation in plant populations growing in patchy heterogeneous landscapes. We study the evolution of a polygenic trait subject to stabilizing selection within populations, but divergent selection between populations. Deviations from linkage equilibrium and Hardy-Weinberg equilibrium make different contributions to genotypic variance depending on the dispersal mode. Local genotypic variance, differentiation between populations and genetic load vary with the rate of gene flow but are similar for seed and pollen dispersal, unless the landscape is very heterogeneous. In this case, genetic load is higher in the case of pollen dispersal, which appears to be due to differences in the distribution of genotypic values before selection.     

Metapopulation models for extinction threshold in spatially correlated landscapes

Simple analytical models assuming homogeneous space have been used to examine the effects of habitat loss and fragmentation on metapopulation size. The models predict an extinction threshold, a critical amount of suitable habitat below which the metapopulation goes deterministically extinct. The consequences of non-random loss of habitat for species with localized dispersal have been studied mainly numerically. In this paper, we present two analytical approaches to the study of habitat loss and its metapopulation dynamic consequences incorporating spatial correlation in both metapopulation dynamics as well as in the pattern of habitat destruction. One approach is based on a measure called metapopulation capacity, given by the dominant eigenvalue of a "landscape" matrix, which encapsulates the effects of landscape structure on population extinctions and colonizations. The other approach is based on pair approximation. These models allow us to examine analytically the effects of spatial structure in habitat loss on the equilibrium metapopulation size and the threshold condition for persistence. In contrast to the pair approximation based approaches, the metapopulation capacity based approach allows us to consider species with long as well as short dispersal range and landscapes with spatial correlation at different scales. The two methods make dissimilar assumptions, but the broad conclusions concerning the consequences of spatial correlation in the landscape structure are the same. Our results show that increasing correlation in the spatial arrangement of the remaining habitat increases patch occupancy, that this increase is more evident for species with short-range than long-range dispersal, and that to be most beneficial for metapopulation size, the range of spatial correlation in landscape structure should be at least a few times greater than the dispersal range of the species.     

Metapopulation persistence and species spread in river networks

River networks define ecological corridors characterised by unidirectional streamflow, which may impose downstream drift to aquatic organisms or affect their movement. Animals and plants manage to persist in riverine ecosystems, though, which in fact harbour high biological diversity. Here, we study metapopulation persistence in river networks analysing stage‐structured populations that exploit different dispersal pathways, both along‐stream and overland. Using stability analysis, we derive a novel criterion for metapopulation persistence in arbitrarily complex landscapes described as spatial networks. We show how dendritic geometry and overland dispersal can promote population persistence, and that their synergism provides an explanation of the so‐called `drift paradox’. We also study the geography of the initial spread of a species and place it in the context of biological invasions. Applications concerning the persistence of stream salamanders in the Shenandoah river, and the spread of two invasive species in the Mississippi‐Missouri are also discussed.     

Metapopulation processes and persistence in remnant water vole populations

We examined the spatial distribution of water vole populations in four consecutive years and investigated whether the regional population processes of extinction, recolonisation and migration influence distribution and persistence. We examined how such regional processes are influenced by spatial variation in habitat quality. In addition, we assessed the relevance of metapopulation concepts for understanding the dynamics of species that deviate from classical metapopulation assumptions and developing conservation measures for them. Populations were patchy and discrete, and the patchy distribution was not static between years. Population turnover occurred even in the absence of predatory mink, which only influenced the network of populations at the end of the study. Most populations were clustered close together in the upper tributaries. Local population persistence was predominantly influenced by population size: large populations were more persistent. Recolonisation rates were influenced by isolation and habitat quality. The isolation estimates which best explained the distribution of water vole populations incorporated straight‐line distances, suggesting water voles disperse overland. The distribution of recolonised sites indicated that dispersing voles actively selected habitat on the basis of its quality. Water voles depart from some of the assumptions made by frequently used metapopulation models. In particular there is no clear binary distinction between suitable and non‐suitable habitat. Accounting for variation in habitat quality before investigating temporal changes in population distribution allowed us to demonstrate that the key metapopulation processes were important. The significance of regional population processes relative to local population processes may have increased in declining, fragmented populations compared to pristine regional populations. We hypothesise that although mink predation is likely to eventually cause regional extinction in many areas, metapopulation processes have delayed this decline. Consequently, conservation measures should take into account mink predation rates and regional population processes, before considering aspects of habitat quality.     

Metapopulation structure and habitat quality in modelling dispersal in the butterfly Iolana iolas

Quantifying dispersal is fundamental to understanding the effects of fragmentation on populations. Although it has been shown that patch and matrix quality can affect dispersal patterns, standard metapopulation models are usually based on the two basic variables, patch area and connectivity. In 2004 we studied migration patterns among 18 habitat patches in central Spain for the butterfly Iolana iolas, using mark–release–recapture methods. We applied the virtual migration (VM) model and estimated the parameters of emigration, immigration and mortality separately for males and females. During parameter estimation and model simulations, we used original and modified patch areas accounting for habitat quality with three different indices. Two indices were based on adult and larval resources (flowers and fruits) and the other one on butterfly density. Based on unmodified areas, our results showed that both sexes were markedly different in their movements and mortality rates. Females emigrated more frequently from patches, but males that emigrated were estimated to move longer daily dispersal distances and suffer higher mortality than females during migration. Males were more likely to emigrate from small than from large patches, but patch area had no significant effect on female emigration. The effects of area on immigration rate and the within-patch mortality were similar in both sexes. Based on modified areas, the estimated parameter values and the model simulation results were similar to those estimated using the unmodified patch areas. One possible reason for the failure to significantly improve the parameter estimates of the VM model is the fact that resource quantity and butterfly population sizes were strongly correlated with patch area. Our results suggest that the standard VM modelling approach, based on patch area and connectivity, can provide a realistic picture of the movement patterns of I. iolas .     

Evolution of dispersal distance

The problem of how often to disperse in a randomly fluctuating environment has long been investigated, primarily using patch models with uniform dispersal. Here, we consider the problem of choice of seed size for plants in a stable environment when there is a trade off between survivability and dispersal range. Ezoe (J Theor Biol 190:287–293, 1998) and Levin and Muller-Landau (Evol Ecol Res 2:409–435, 2000) approached this problem using models that were essentially deterministic, and used calculus to find optimal dispersal parameters. Here we follow Hiebeler (Theor Pop Biol 66:205–218, 2004) and use a stochastic spatial model to study the competition of different dispersal strategies. Most work on such systems is done by simulation or nonrigorous methods such as pair approximation. Here, we use machinery developed by Cox et al. (Voter model perturbations and reaction diffusion equations 2011) to rigorously and explicitly compute evolutionarily stable strategies.     

From dispersal to landscapes: progress in the understanding of population dynamics

The development of our understanding of population dynamics over the past 50 years is reviewed from a personal perspective. An early emphasis on population vital rates was superceded by recognition of the importance of the specific community context of focal populations, and most recently has in turn been enriched by a landscape perspective. Certain basic principles are outlined including the value of a systems context for population analyses, the power of a dual mechanistic and contextual perspective, and the inevitability of density control in a finite biosphere. Numbers are determined by the balance of two complex parameters:p — the per capita growth promoting (enhancing) forces, ands — the per capita growth suppressing forces. Multiple factor explanations of demographic behavior are therefore to be expected, as well as temporal and spatial variations in them. An appreciation for the potential role of dispersal as a population vital rate led to the development of metapopulation theory. A renewed understanding of the role of community context in population dynamics provoked the realization that a multi-factor approach was required. This in turn allowed us to reconcile the reality of local demographic complexity with global generalizations. Finally, the introduction of landscape ecology into demographic thinking added many new insights. It is now appreciated that a spatially explicit mosaic of habitat patches, edge effects, corridors, and even the proportion of favorable to marginal habitats can all be critically important factors in influencing population dynamics.     

Evolution of reduced dispersal mortality and 'fat-tailed' dispersal kernels in autocorrelated landscapes

Models describing the evolution of dispersal strategies have mostly focused on the evolution of dispersal rates. Taking trees as a model for organisms with undirected, passive dispersal, we have developed an individual-based, spatially explicit simulation tool to investigate the evolution of the dispersal kernel, P(r), and its resulting cumulative seed-density distribution, D(r). Simulations were run on a variety of fractal landscapes differing in the fraction of suitable habitat and the spatial autocorrelation. Starting from a uniform D(r), evolution led to an increase in the fraction of seeds staying in the home cell, a reduction of the dispersal mortality (arrival in unsuitable habitat), and the evolution of 'fat-tailed' D(r) in autocorrelated landscapes and approximately uniform D(r) in random landscapes. The evolutionary process was characterized by long periods of stasis with a few bouts of rapid change in the dispersal rate.     

Metapopulation extinction in fragmented landscapes: using bacteria and protozoa communities as model ecosystems

Extinction is notoriously difficult to study because of the long timescales involved and the difficulty in ascertaining that extinction has actually occurred. The effect of habitat subdivision, or fragmentation, on extinction risk is even harder to study, as it requires copious replication of habitat patches on large spatial scales and control of area effects between treatments. I used simple small-scale communities of bacteria and protozoa to study extinction in response to habitat loss and habitat fragmentation. I studied several different community configurations, each with three trophic levels. Unlike most metapopulation studies (experimental as well as theoretical), which have tended to deal with inherently unstable species interactions, I deliberately used community configurations that were persistent in large stock cultures. I recorded the time to extinction of the top predator in single habitat patches of different sizes and in fragmented systems with different degrees of subdivision but the same amount of available habitat. Habitat loss reduced the time to extinction of isolated populations. Fragmented systems went extinct sooner than corresponding unfragmented (continuous) systems of the same overall size. Unfragmented populations persisted longer than fragmented systems (metapopulations) with or without dispersal corridors between subpopulations. In fact, fragmented systems where the fragments were linked by dispersal corridors went extinctly significantly sooner than those where subpopulations were completely isolated from each other. If these results extend to more "natural" systems, it suggests a need for caution in management programs that emphasize widespread establishment of wildlife corridors in fragmented landscapes.