On the other hand, larval dispersal is often modelled as an advection—diffusion process e. This may be appropriate for assessing the long-term pattern of larval transport but will not describe transport for a single spawning season Siegel et al. Most nearshore marine species have a brief spawning window of days to at most a few months, so the number of statistically independent dispersal paths emerging from a source location will be small Siegel et al.
Mitarai et al. The simulated patterns of larval connectivity were spatially heterogeneous for a single spawning season and were highly variable among years. They also showed that the statistical properties of the connectivity patterns could be captured by caricaturing the process as a handful of successful dispersal events, where a single event links a contiguous group of source locations with a contiguous group of destination locations Mitarai et al.
The number and size of these events, as well as the mean and variance of distance travelled by the larvae in an event, depend on the characteristics of the flow such as mean eddy size , the length of the spawning season, and the duration of the pelagic dispersal period.
Siegel et al. In this study, we demonstrate that stochastic dispersal, as experienced by many nearshore marine organisms, can promote species coexistence.
We analyse a simple competition model, loosely based on life history characteristics typical of shallow-dwelling reef fishes such as kelp rockfish Sebastes atrovirens ; Love et al.
We have deliberately excluded priority effects, spatial or temporal heterogeneity in environmental quality, or life-history tradeoffs — while these processes may occur in rockfish, they are already known to promote coexistence, and we want to focus on the role of stochastic dispersal. Nevertheless, we find coexistence of two similar competing species that could not coexist in a non-spatial model if their spawning seasons do not perfectly overlap.
We model two competing species distributed along a linear coastline, divided into evenly spaced sites of suitable habitat. Adults remain within a site and all competitive interactions are local. For simplicity, we abstract away many aspects of life history age and size structure of adults, adult competition, spatial heterogeneity in habitat quality. Competition occurs among settlers at a location, creating density-dependent recruitment:.
We assume that the inter- and intraspecific interaction strengths are the same for both species, because larvae of co-settling rockfish are often morphologically and ecologically indistinguishable Wilson et al.
Larval mortality during dispersal from predation, starvation, and being swept out to sea depends on the time in the plankton, which can span a range of days to months depending on the species; mortality variation due to the time required to disperse from x to y can be incorporated in D.
The larval production term includes the average larval mortality and represents the expected number of settlers produced by an adult. Therefore, the expected value of is equal to 1. We allow the two competing species to differ in two ways. All other parameters are identical between the two species, and are constant in space and time; hence, we drop the species-specific subscripts.
While species A has a fitness advantage over species B, the two species are competitively equivalent: the relative frequency of recruits at a location is the same as the relative frequency of settlers. As the two species have identical average dispersal characteristics, species B has no advantage over species A, and there is no opportunity for coexistence via a tradeoff.
The parameters f , a and m jointly control the intensity of density dependence when the population is at carrying capacity; in the analyses below we vary f to examine the effects of density dependence higher fecundity produces more settlers, so that maintaining the same total recruitment at equilibrium requires more intense competition.
We assumed that spawning lasts 30 days, and larvae can settle if they encounter the coastline between the ages of 20 and 40 days. The connectivity patterns generated by the ROMS model are spatially heterogeneous and temporally stochastic; Fig. Larval release and settlement are spatially correlated, but the specific locations of these events change depending on the exact realization of the mesoscale 20— km flow field Mitarai et al.
These eddies collect larvae released from the nearshore over a large spatial scale and transport and deliver them as settlers in a cohesive unit. However, in any given year, connectivity is patchy and the patterns vary substantially from year to year. Simulations of realized dispersal in the ROMS model a—c and the packet model d—f. Each of the three panels in each row represents a different year and the color scale represents the number of larvae dispersing from a given source to a given destination.
The spatial variance in the packet model connectivity matrix is 0. The ROMS model runs too slowly to incorporate into a population model. Instead, we take advantage of previous work that developed a much simpler model that captures the general statistical patterns of larval dispersal in the turbulent ocean Mitarai et al.
The number of packets in a given year is. The destination location, x k , for the k th packet is selected randomly from within the domain while its source location, y k , is drawn from a normal distribution representing the long-term mean dispersal kernel.
Connectivity matrices are then modelled based upon the number of packets between a given source and destination spread over the eddy scale, r :. The packet model connectivity patterns are more artificial-looking than those from the ROMS simulations Fig. The patterns of variability within a year will also turn out to be important: how does realized connectivity vary between two seasons that do not overlap, or only partially overlap?
For each site, we define an integrated measure of connection in a given year by summing D across all sources:. This spatial covariance increases nearly linearly with the overlap in spawning seasons with similar patterns arising from the ROMS simulations and the packet model Fig. Thus, the packet model provides a sound approximation to hydrodynamically realistic dispersal for use in spatially explicit population models.
In the spatially explicit model with diffusive dispersal, the high-productivity species species A drives the low-productivity species species B to extinction Fig. In contrast, the low-productivity species can persist when the two species disperse according to independent realizations of the packet model Fig.
Spatiotemporal patterns in adults of both species are patchy Fig. Despite the spatial variability, mean abundances are relatively constant once the species approach their equilibria Fig. Left: Mean population size of both species through time using diffusive dispersal a and packet model dispersal with no overlap in spawning b.
Right: Spatio-temporal patterns in population size using packet model dispersal for species A c and species B d for a year time span and over the centre km of the domain. Coexistence in this model requires some decorrelation in settlement: the two species must not have exactly the same realized dispersal kernel, which would arise if they had exactly the same spawning season and competency window.
Coexistence depends on the overlap in spawning seasons Fig. As the amount of overlap increases, the equilibrium abundance of species B declines until it is effectively extinct, above an overlap of 25 days. Mean percentage of the total population size for each species after years mean of 50 simulations over a range of overlap in spawning from none to complete.
This model contains several processes and emergent patterns that might contribute to coexistence. For example, larval settlement is highly aggregated i. This aggregation, together with the spatial autocorrelation in settlement incorporated in the packet model, leads to substantial spatio-temporal patterning in adult abundance Fig.
Finally, the combination of intraspecific variability and imperfect interspecific correlation in settlement suggests that a storage effect may be acting.
To tease these apart we turn to a simpler model, containing only the latter mechanism. To focus on the role of dispersal variability in promoting coexistence, we develop a spatially implicit model that strips away the potentially confounding factors discussed above. First, we eliminate intra- and interspecific patterns in adult density, forcing adult density to be homogeneous at the end of each time step:.
This is not meant to be a biologically realistic approximation although it could be achieved by assuming a high rate of adult movement ; rather, we are artificially intervening to ensure that any remaining coexistence is not due to adult spatial patterning.
Second, we eliminate the aggregation and spatial autocorrelation in settlement, focusing on the simple effects of spatial variances regardless of magnitude and interspecific correlations.
Because adult densities are homogeneous, the number of settlers at location x is. We do not incorporate spatial autocorrelation, and we assume that all higher moments of dispersal variation are zero. As species A has higher fitness, the coexistence criterion is that species B must be able to increase when it is at low density Chesson This means that average per-capita recruitment must exceed adult mortality:. As species A is at equilibrium, its expected per-capita recruitment equals its mortality:.
We now estimate the expected recruitment of both species, keeping terms up to second order in the Taylor expansion around the mean number of settlers:. Substituting eqns 15 and 16 into inequality 10 gives the coexistence condition in terms of the ratio of productivity between the species:. This can be evaluated using estimates of the dispersal variance and covariance from the packet model.
The higher the correlation in dispersal, the more demographically similar the species need to be in order to coexist Fig.
The range of fitness differences over which coexistence is possible expands with increasing intensity of competition f A ; Fig. There is no limiting similarity in this model: coexisting species can have arbitrarily small differences in dispersal patterns, as long as the demographic differences are also small. This can be understood by recognizing that the variance in settler abundance is proportional to the square of mean adult abundance, so reducing the mean adult abundance disproportionally reduces the spatial variability in the competitive environment available for the invading species to exploit.
Coexistence thresholds from the spatially implicit model, relating correlation in settlement and the fecundity ratio of species B to species A. Coexistence occurs to the right of the lines. Panel a varies the intensity of density dependence by changing the fecundity of Species A. Panel b varies the variance in settlement. Does the spatially implicit model, with its reduced set of features and processes, capture the coexistence properties of the spatially explicit model?
Except when the density dependence is strongest, the spatially implicit model predicts coexistence in the spatially explicit model almost perfectly Fig. This provides strong evidence that coexistence in the spatially explicit model is predominantly produced by spatial variability in settlement combined with some level of settlement decorrelation between species; turbulent dispersal is simply providing a means to achieve appropriate settlement statistics.
The other phenomena in the simulation model spatial patterns in adult density, aggregation in settlement are quantitatively and qualitatively irrelevant to coexistence.
Coexistence thresholds estimated from the spatially explicit and spatially implicit models. We analyzed the effect of the sex of resident juveniles on settlement decisions of dispersers with chi-square tests.
To investigate the cost of dispersal, we compared the mortality rates of radio-tagged dispersers with radio-tagged resident juveniles retained juveniles and immigrants during the same period between mid-June and mid-August. Mortality rates between dispersers, retained offspring, and immigrants were compared using a GLMM in Genstat binary error distribution, logit link to assess the effect of individual category on mortality.
In addition to year, we included individual identity as a random factor in this model because some individuals were observed both as dispersers and as immigrants. The costs of immigrants to breeders were assessed by analyzing the effect that immigrants had on the breeding success and nestling growth.
Breeders are hostile toward immigrants and regularly displace or chase them, especially when immigrants try to approach the nest during the breeding season Ekman et al. An increased investment in aggression could reduce the time available for parental care and hence cause a reduction of nestling growth or breeding success Eggers We analyzed the effect of the number of immigrants on a territory on the weight of nestlings using a stepwise forward structured general linear model normal error distribution, identity link.
Tarsus length and time of the day were entered into the model before entering the number of immigrants. We also included habitat quality and the number of retained offspring in the model. Territory and year were included as random factors to control for repeated measurements of the same brood.
With a similar model GLMM in Genstat; Poisson error distribution, logarithm link , we estimated the effect of the number of immigrants on the number of fledglings produced on a territory. During 16 of these encounters, retained juveniles chased the disperser for an average of All encounters resulted in the dispersers leaving the territory and continuing dispersal.
Older group members retained offspring, immigrants, breeders were less aggressive, and their response depended on the presence of retained juveniles Table 2.
The aggressive behavior of residents was specifically directed toward dispersers and not just a general response toward juveniles. Residents never displaced or chased resident juveniles of neighboring groups Table 2 ; Figure 1. The number of observations for each group is shown in the bottom of the bars. Territory identity and year were included as random variable in model to control multiple observations in same group. Despite that unoccupied territories were frequently available Table 1 , no disperser settled in an empty territory.
Instead, all dispersers immigrated into existing groups and remained in the groups where they initially settled. Only 7 dispersers 4. We tested experimentally whether this low rate of direct occupation of breeding openings was a result of a difference in competitive abilities between dispersers and resident individuals.
When examining individual settlement decisions, the only factor affecting the number of dispersers settling on a territory was the number of retained juveniles on a territory Table 3. Dispersers generally settled on territories without retained juveniles out of immigrants; Figure 2.
Only 13 immigrants settled on a territory with one retained juvenile, whereas just 2 immigrants settled in a group with 2 retained juveniles. No other factor habitat quality, breeder phenotype, number of older group members affected settlement decisions.
By definition, groups of only 2 group members never contained any retained juveniles. No dispersers settled in groups of 5 individuals. GLMM of the factors associated with the number of juvenile immigrants on a territory. Effects indicate direction of relationship and are presented after setting the mean of the covariate to zero. Although immigrants avoided settling on a territory where retained juveniles were present, 67 immigrants settled together with other immigrants on a territory.
Immigrants settled preferable in groups with short queues avoiding groups where same-sex juveniles were present. The presence of immigrants during the breeding season March—May had a negative effect on nestling weight. Nestlings in groups with immigrants had a significantly lower weight than nestlings in groups without immigrants when controlling for tarsus length Table 4. Neither the number of retained offspring present during reproduction nor habitat quality affected nestling condition.
The assumption that behavioral interactions between dispersers and residents largely determine dispersal and settlement decisions is reasonably well established in the literature on dispersal Krebs ; Rosenberry et al. However, this process has remained poorly investigated Clobert et al.
We found that retained Siberian jay juveniles aggressively approached all dispersers and chased them off their territory, thereby preventing dispersers from settling and forcing them to move on. As a consequence of this aggression toward dispersers, immigrants settled mainly in groups where reproduction had previously failed and hence where no retained juveniles were present. Aggression of residents toward dispersers could be a widespread behavior and has been suggested to prevent dispersers from settlement in other species Holekamp and Smale ; Nunes et al.
Moreover, dispersal and settlement decisions in Siberian jays affect fitness of both dispersers inability to choose high-quality patches, high winter mortality; Ekman et al. In the Siberian jay and its North American congener species, the gray jay Perisoreus canadensis Strickland , natal juvenile dispersal is a consequence of sibling rivalry within broods reflecting social dominance after fledging Ekman et al.
The individuals that evict subordinate brood members from the natal territory are the same individuals that evict dispersers from their territory. This is similar to the behavioral mechanism of group eviction in anemonefish and coral-dwelling goby Paragobiodon xanthosomus where aggression is primarily directed from a more dominant individual toward its immediate subordinate Fricke and Fricke ; Wong et al.
The lowest ranking group members are thus in these systems in control of settlement decisions Buston b ; Wong et al. In contrast, in many other group-living species, breeders are in control of subordinate group membership green jay Cyanocorax yncas : Gayou ; primates: Pusey and Packer ; superb fairy-wren Malurus cyaneus : Mulder ; daffodil cichlid: Heg et al. Most models on dispersal consider only the later case where breeders determine group membership of nonbreeding group members Clobert et al.
All dispersers in Siberian jays settled on existing territories, despite that empty territories were available in most years. Further studies are needed to determine whether this is an effect of dispersers using conspecific attraction as a cue to recognize appropriate habitat Stamps or whether benefits of group living outweigh settlement in an empty territory.
The settlement patterns of Siberian jays have drastic consequences for the population dynamic of this species. In contrast to the long distances juvenile dispersers move Ekman et al. Reoccupation of isolated unoccupied large areas of suitable habitat will therefore take a long time.
This may have a negative effect on future population size because modern forestry often results in serious fragmentation to the landscape Hansson and jays abandon intensively managed areas Griesser et al. Queue length has been demonstrated to be a a poor predictor of settlement decisions across species.
Whereas dispersing superb fairy-wrens settle preferably on territories with long queues Cockburn et al. In our study, not only queue length and in particular the presence of retained juveniles but also disperser sex affected settlement decisions.
Immigration of same-sex dispersers is costly for retained juveniles because they directly compete over access to future breeding openings. Not surprisingly, we observed few jay groups containing 2 same-sex juveniles. The breeder removal experiment demonstrated that in Siberian jays dispersers and retained juveniles cannot compete with older individuals retained offspring, immigrants, widowed breeders over breeding openings. This may explain why the number of older retained offspring and immigrants in a group had no influence on the number of immigrants settling in a territory.
Queues in Siberian jay seem thus to be stable, which stands in agreement with many empirical and theoretical studies that report or predict stable queues Kokko and Johnstone ; Buston ; but see Kokko and Ekman Dispersing jays did not have a higher mortality rate compared with same-aged residents during the dispersal period.
Despite the importance of dispersal costs for dispersal models Clobert et al. Increased mortality of dispersers may result from moving through unfamiliar space Greenwood and Harvey , Yoder et al. In our study, immigrants ended up queuing for breeding openings in low-quality territories where prior reproduction had failed because of the despotic behavior of retained juveniles.
This suggests that social constraints confine dispersers to settle on low-quality territories, affecting their future fitness Ekman et al. Breeders are dominant over immigrant group members, and thus, the initial costs of letting a disperser settle on their territory are low Ekman and Sklepkovych Because immigrants are always unrelated to the breeders, opposite-sex immigrants can immediately replace dead partners Ekman and Griesser , and breeders are only aggressive toward same-sex immigrants but tolerant toward opposite-sex immigrants Ekman and Sklepkovych However, immigrants still induce some cost to the breeders through increased levels of aggression during the breeding season Ekman et al.
Suboptimal growth during the nestling period reduces first winter survival Griesser et al. The presence of immigrants thus poses also a small fitness cost to retained offspring through lowering the fitness of their siblings. Recent research has identified different proximate mechanisms influencing dispersal distance ranging from morphological traits, physiological traits Nunes et al.
Many of these traits are heritable Roff and Fairbairn , and singe alleles have been linked to dispersal rates of individuals Haag et al. However, we still lack knowledge about the behavioral decisions leading to a specific dispersal path of an individual Macdonald and Johnson ; Doerr and Doerr Our results demonstrate the importance of social interactions between residents and dispersers on both the dispersal process and settlement decisions.
Given that group living is widespread in vertebrates, an understanding of social interactions on dispersal decisions will be important to link the proximate, behavioral decisions with the ultimate, evolutionary consequences of dispersal to bring us a step closer to a comprehensive dispersal framework that links theory with empirical data. The Swedish Research Council to J.
However, these areas may be beyond the dispersal capacity of many species. Ultimate causes of dispersal can be explained by avoidance of inbreeding and inbreeding depression.
Small, isolated populations can become inbred and result in decreased fitness, but dispersal can counteract these negative effects. Additionally, dispersal can reduce competition for resources and mates, thereby increasing individual fitness. In some situations, these ultimate causes will result in sex-biased dispersal. For example, mammals typically exhibit male-biased dispersal, and birds typically exhibit female-biased dispersal.
These dispersal strategies result mostly from males attempting to increase their access to females male-biased dispersal and in female-biased dispersal systems in birds from male resource defense female-biased dispersal in birds results Greenwood Despite the perceived benefits of dispersal, there can be costs.
First and foremost, there is a greater mortality risk during dispersal due to increased energy expenditure, unfamiliar habitat, or predation risk e.
Second, dispersers may suffer reduced survival or reproductive success because of unfamiliarity with the new environment and the inability to acquire sufficient resources, resulting in decreased adaptive ability to the new habitat. Dispersal affects organisms at individual, population, and species levels. Survival, growth, and reproduction at the level of individuals are intimately tied to both the distance and frequency of dispersal, factors which are typically mediated by aspects of local resource availability.
At the population level, patterns of emigration and immigration within and among habitat patches associated with local population density, among other factors, drive temporal and spatial cycles of colonization and extinction. The form of such movements, such as stepping-stone versus one-way migration, ultimately determines the genetic structure of populations, wherein genetic differentiation is directly proportional to the amount of gene flow among populations.
For populations exhibiting frequent dispersal, ongoing gene flow within and among populations results in those populations becoming genetically similar to one another and ultimately evolving as a single unit. Finally, over evolutionary time frames, a lack of dispersal among populations impacts organisms at the species level. If dispersal between populations ceases, these newly isolated populations accumulate novel genetic attributes via genetic drift or natural selection potentially leading to local adaptation.
Insurmountable landscape features, such as mountains and rivers, typically drive such processes, and in cases where genetic differentiation persists even after dispersal between formerly isolated populations could resume, such entities can then be designated as separate species Figure 3. Figure 3: Phylogenetic relationships of hypothetical populations that became isolated via dispersal Uppercase letters represent taxa, roman numerals represent geographic areas, black arrows represent dispersal events.
All rights reserved. Species exhibit geographic distributions that are constrained by a range of environmental variables — outside of which individuals may experience reduced survival and reproduction due to physical and physiological constraints.
For example, species are often accustomed to particular temperature ranges, and dispersal to regions with temperatures outside those ranges reduces fitness. Additionally, resources necessary for population persistence may be insufficient at range edges and outside the range.
Physical barriers to dispersal consist of landscape features that prevent organisms from relocating. Mountains, rivers, and lakes are examples of physical barriers that can limit a species' distribution. Anthropogenic barriers, like roads, farming, and river dams, also function as impediments to movement. It has been suggested that anthropogenic barriers are the most serious threats to dispersal. These barriers can effectively divide up a species' range into isolated fragments, and dispersal from one habitat patch to another can prove difficult.
Creating dispersal corridors has been suggested as a means to maintain connectivity between habitat patches. For example, Banff National Park in Alberta, Canada, contains 22 underpasses and 2 overpasses to facilitate wildlife dispersal within the park across a busy four-lane highway the Trans-Canada Highway. Similarly, wildlife crossings, specifically designed for Florida panthers, were constructed along a forty-mile stretch of Interstate 75 in Florida.
Corridors are not just for large mammals either: Salamanders have also benefited from miniature underpasses to facilitate dispersal. Additionally, recent research has focused on using modeling techniques to analyze available habitat to designate potential dispersal pathways for species whose ranges have been fragmented Figure 4. Source populations in the West were as follows: A.
Badlands, ND; B. Black Hills, SD; C. Kimble County, TX. Anabrus simplex with radio transmitters attached see Lorch et al. Direct methods can be somewhat easier to use in larger animals simply because tracking the smallest organisms e. However, tracking devices are becoming increasingly more advanced and useful in small organisms Figure 5. Interpretation of results from direct measurement can sometimes prove difficult though. Low accuracy of spatial position, disproportionate mortality of marked individuals, labor intensity, and high costs are all deterrents to using direct measurement methods.
In contrast to direct methods, indirect methods infer the degree of dispersal without actually having to observe the dispersal movement.
Typically, indirect methods involve utilizing molecular markers to measure gene flow and deduce dispersal patterns based on within and among population genetic differences. Specifically, the differences in allele or genotype frequencies resulting from gene flow between populations reveal patterns and levels of dispersal. Indirect methods are increasingly being used to infer dispersal because of the difficulties involved with direct measurement. Human activities have facilitated and impeded dispersal in many ways.
As stated previously, anthropogenic barriers in the form of human development have disrupted natural dispersal patterns in a variety of species. Conversely, humans have also facilitated dispersal, both deliberately and accidentally. A common inadvertent way organisms have been dispersed is through their transport in the ballast water of ships.
Ships emptying ballast water may release foreign organisms. For example, zebra mussels, a freshwater mollusk native to the lakes of southeast Russia, were accidentally introduced into the Great Lakes of North America where they have caused major economic problems by clogging water treatment and power plants through ballast water discharge.
As a result of the potential for introduction of non-native organisms via ballast water, new standards have been proposed for ballast tank cleaning. Humans have also transported organisms to areas outside their native ranges for deliberate reasons. The seeds of attractive plants native to areas outside North America are routinely used in gardens and have the capacity to disperse to wild areas if conditions are suitable e. Also, bighead and silver carp originating from China were introduced to catfish farm ponds in the United States to control algal growth.
Fish accidentally escaped from these ponds and have subsequently colonized the Mississippi, Missouri, Illinois and Ohio rivers where they have had significant negative impact on the native fauna Figure 6.
Dispersal is a common process undertaken by individuals at different stages of the life cycle and in response to various factors. Morphological adaptations make dispersal achievable but with varying degrees of success due to anthropogenic and natural barriers.
These barriers modify the level of dispersal and consequently exert effects on population dynamics and genetic structure. As environments are altered, through stochastic events and global climate change, it will become increasingly important to assess how such changes will affect dispersal at the individual, population, and species levels.
Avise, J. Phylogeography: The History and Formation of Species. Freeze, M. North American Journal of Fisheries Management 2 , — doi Johnson, C. Mortality risk increases with dispersal distance in American martens. Larue, M. Modelling potential dispersal corridors for cougars in midwestern North America using least-cost path methods. Ecological Modelling , — doi Lorch, P. Radiotelemetry reveals differences in individual movement patterns between outbreak and non-outbreak Mormon cricket populations.
Ecological Entomology 30 , — doi Mate, B. Satellite-monitored movements of the northern right whale.
0コメント