Theoretical ecology

Exponential growth

The most basic way of modeling population dynamics is to assume that the rate of growth of a population depends only upon the population size at that time and the per capita growth rate of the organism. In other words, if the number of individuals in a population at a time t, is N(t), then the rate of population growth is given by:

${\displaystyle {\frac {dN(t)}{dt}}=rN(t)}$

where r is the per capita growth rate, or the intrinsic growth rate of the organism. It can also be described as r = b-d, where b and d are the per capita time-invariant birth and death rates, respectively. This first order linear differential equation can be solved to yield the solution

${\displaystyle N(t)=N(0)\ e^{rt}}$,

a trajectory known as Malthusian growth, after Thomas Malthus, who first described its dynamics in 1798. A population experiencing Malthusian growth follows an exponential curve, where N(0) is the initial population size. The population grows when r > 0, and declines when r < 0. The model is most applicable in cases where a few organisms have begun a colony and are rapidly growing without any limitations or restrictions impeding their growth (e.g. bacteria inoculated in rich media).

Logistic growth

The exponential growth model makes a number of assumptions, many of which often do not hold. For example, many factors affect the intrinsic growth rate and is often not time-invariant. A simple modification of the exponential growth is to assume that the intrinsic growth rate varies with population size. This is reasonable: the larger the population size, the fewer resources available, which can result in a lower birth rate and higher death rate. Hence, we can replace the time-invariant r with r’(t) = (b –a*N(t)) – (d + c*N(t)), where a and c are constants that modulate birth and death rates in a population dependent manner (e.g. intraspecific competition). Both a and c will depend on other environmental factors which, we can for now, assume to be constant in this approximated model. The differential equation is now:^[16]

${\displaystyle {\frac {dN(t)}{dt}}=((b-aN(t))-(d-cN(t)))N(t)}$

This can be rewritten as:^[16]

${\displaystyle {\frac {dN(t)}{dt}}=rN(t)\left(1-{\frac {N}{K}}\right)}$

where r = b-d and K = (b-d)/(a+c).

The biological significance of K becomes apparent when stabilities of the equilibria of the system are considered. The constant K is the carrying capacity of the population. The equilibria of the system are N = 0 and N = K. If the system is linearized, it can be seen that N = 0 is an unstable equilibrium while K is a stable equilibrium.^[16]

Structured population growth

Another assumption of the exponential growth model is that all individuals within a population are identical and have the same probabilities of surviving and of reproducing. This is not a valid assumption for species with complex life histories. The exponential growth model can be modified to account for this, by tracking the number of individuals in different age classes (e.g. one-, two-, and three-year-olds) or different stage classes (juveniles, sub-adults, and adults) separately, and allowing individuals in each group to have their own survival and reproduction rates. The general form of this model is

${\displaystyle \mathbf {N} _{t+1}=\mathbf {L} \mathbf {N} _{t}}$

where N_t is a vector of the number of individuals in each class at time t and L is a matrix that contains the survival probability and fecundity for each class. The matrix L is referred to as the Leslie matrix for age-structured models, and as the Lefkovitch matrix for stage-structured models.^[17]

If parameter values in L are estimated from demographic data on a specific population, a structured model can then be used to predict whether this population is expected to grow or decline in the long-term, and what the expected age distribution within the population will be. This has been done for a number of species including loggerhead sea turtles and right whales.^[18][19]

Predator–prey interaction

Predator–prey interactions exhibit natural oscillations in the populations of both predator and the prey.^[21] In 1925, the American mathematician Alfred J. Lotka developed simple equations for predator–prey interactions in his book on biomathematics.^[22] The following year, the Italian mathematician Vito Volterra, made a statistical analysis of fish catches in the Adriatic^[23] and independently developed the same equations.^[24] It is one of the earliest and most recognised ecological models, known as the Lotka-Volterra model:

${\displaystyle {\frac {dN(t)}{dt}}=N(t)(r-\alpha P(t))}$

${\displaystyle {\frac {dP(t)}{dt}}=P(t)(c\alpha N(t)-d)}$

where N is the prey and P is the predator population sizes, r is the rate for prey growth, taken to be exponential in the absence of any predators, α is the prey mortality rate for per-capita predation (also called ‘attack rate’), c is the efficiency of conversion from prey to predator, and d is the exponential death rate for predators in the absence of any prey.

Volterra originally used the model to explain fluctuations in fish and shark populations after fishing was curtailed during the First World War. However, the equations have subsequently been applied more generally.^[25] Other examples of these models include the Lotka-Volterra model of the snowshoe hare and Canadian lynx in North America,^[26] any infectious disease modeling such as the recent outbreak of SARS ^[27] and biological control of California red scale by the introduction of its parasitoid, Aphytis melinus .^[28]

A credible, simple alternative to the Lotka-Volterra predator–prey model and their common prey dependent generalizations is the ratio dependent or Arditi-Ginzburg model.^[29] The two are the extremes of the spectrum of predator interference models. According to the authors of the alternative view, the data show that true interactions in nature are so far from the Lotka–Volterra extreme on the interference spectrum that the model can simply be discounted as wrong. They are much closer to the ratio-dependent extreme, so if a simple model is needed one can use the Arditi–Ginzburg model as the first approximation.^[30]

Host–pathogen interaction

The second interaction, that of host and pathogen, differs from predator–prey interactions in that pathogens are much smaller, have much faster generation times, and require a host to reproduce. Therefore, only the host population is tracked in host–pathogen models. Compartmental models that categorize host population into groups such as susceptible, infected, and recovered (SIR) are commonly used.^[31]

Host–parasitoid interaction

The third interaction, that of host and parasitoid, can be analyzed by the Nicholson–Bailey model, which differs from Lotka-Volterra and SIR models in that it is discrete in time. This model, like that of Lotka-Volterra, tracks both populations explicitly. Typically, in its general form, it states:

${\displaystyle N_{t+1}=\lambda \ N_{t}\ [1-f(N_{t},P_{t})]}$

${\displaystyle P_{t+1}=c\ N_{t}\ f(N_{t},p_{t})}$

where f(N_t, P_t) describes the probability of infection (typically, Poisson distribution), λ is the per-capita growth rate of hosts in the absence of parasitoids, and c is the conversion efficiency, as in the Lotka-Volterra model.^[21]

Competition and mutualism

In studies of the populations of two species, the Lotka-Volterra system of equations has been extensively used to describe dynamics of behavior between two species, N₁ and N₂. Examples include relations between D. discoiderum and E. coli,^[32] as well as theoretical analysis of the behavior of the system.^[33]

${\displaystyle {\frac {dN_{1}}{dt}}={\frac {r_{1}N_{1}}{K_{1}}}\left(K_{1}-N_{1}+\alpha _{12}N_{2}\right)}$

${\displaystyle {\frac {dN_{2}}{dt}}={\frac {r_{2}N_{2}}{K_{2}}}\left(K_{2}-N_{2}+\alpha _{21}N_{1}\right)}$

The r coefficients give a “base” growth rate to each species, while K coefficients correspond to the carrying capacity. What can really change the dynamics of a system, however are the α terms. These describe the nature of the relationship between the two species. When α₁₂ is negative, it means that N₂ has a negative effect on N₁, by competing with it, preying on it, or any number of other possibilities. When α₁₂ is positive, however, it means that N₂ has a positive effect on N₁, through some kind of mutualistic interaction between the two. When both α₁₂ and α₂₁ are negative, the relationship is described as competitive. In this case, each species detracts from the other, potentially over competition for scarce resources. When both α₁₂ and α₂₁ are positive, the relationship becomes one of mutualism. In this case, each species provides a benefit to the other, such that the presence of one aids the population growth of the other.

See Competitive Lotka–Volterra equations for further extensions of this model.

Neutral theory

Unified neutral theory is a hypothesis proposed by Stephen P. Hubbell in 2001.^[20] The hypothesis aims to explain the diversity and relative abundance of species in ecological communities, although like other neutral theories in ecology, Hubbell's hypothesis assumes that the differences between members of an ecological community of trophically similar species are "neutral," or irrelevant to their success. Neutrality means that at a given trophic level in a food web, species are equivalent in birth rates, death rates, dispersal rates and speciation rates, when measured on a per-capita basis.^[34] This implies that biodiversity arises at random, as each species follows a random walk.^[35] This can be considered a null hypothesis to niche theory. The hypothesis has sparked controversy, and some authors consider it a more complex version of other null models that fit the data better.

Under unified neutral theory, complex ecological interactions are permitted among individuals of an ecological community (such as competition and cooperation), providing all individuals obey the same rules. Asymmetric phenomena such as parasitism and predation are ruled out by the terms of reference; but cooperative strategies such as swarming, and negative interaction such as competing for limited food or light are allowed, so long as all individuals behave the same way. The theory makes predictions that have implications for the management of biodiversity, especially the management of rare species. It predicts the existence of a fundamental biodiversity constant, conventionally written θ, that appears to govern species richness on a wide variety of spatial and temporal scales.

Hubbell built on earlier neutral concepts, including MacArthur & Wilson's theory of island biogeography^[20] and Gould's concepts of symmetry and null models.^[34]

Biogeography

Biogeography is the study of the distribution of species in space and time. It aims to reveal where organisms live, at what abundance, and why they are (or are not) found in a certain geographical area.

Biogeography is most keenly observed on islands, which has led to the development of the subdiscipline of island biogeography. These habitats are often a more manageable areas of study because they are more condensed than larger ecosystems on the mainland. In 1967, Robert MacArthur and E.O. Wilson published The Theory of Island Biogeography. This showed that the species richness in an area could be predicted in terms of factors such as habitat area, immigration rate and extinction rate.^[36] The theory is considered one of the fundamentals of ecological theory.^[37] The application of island biogeography theory to habitat fragments spurred the development of the fields of conservation biology and landscape ecology.^[38]

r/K-selection theory

Main article: r/K selection

A population ecology concept is r/K selection theory, one of the first predictive models in ecology used to explain life-history evolution. The premise behind the r/K selection model is that natural selection pressures change according to population density. For example, when an island is first colonized, density of individuals is low. The initial increase in population size is not limited by competition, leaving an abundance of available resources for rapid population growth. These early phases of population growth experience density-independent forces of natural selection, which is called r-selection. As the population becomes more crowded, it approaches the island's carrying capacity, thus forcing individuals to compete more heavily for fewer available resources. Under crowded conditions, the population experiences density-dependent forces of natural selection, called K-selection.^[39][40]

Niche theory

Metapopulations

Spatial analysis of ecological systems often reveals that assumptions that are valid for spatially homogenous populations – and indeed, intuitive – may no longer be valid when migratory subpopulations moving from one patch to another are considered.^[42] In a simple one-species formulation, a subpopulation may occupy a patch, move from one patch to another empty patch, or die out leaving an empty patch behind. In such a case, the proportion of occupied patches may be represented as

${\displaystyle {\frac {dp}{dt}}=mp(1-p)-ep}$

where m is the rate of colonization, and e is the rate of extinction.^[43] In this model, if e < m, the steady state value of p is 1 – (e/m) while in the other case, all the patches will eventually be left empty. This model may be made more complex by addition of another species in several different ways, including but not limited to game theoretic approaches, predator–prey interactions, etc. We will consider here an extension of the previous one-species system for simplicity. Let us denote the proportion of patches occupied by the first population as p₁, and that by the second as p₂. Then,

${\displaystyle {\frac {dp_{1}}{dt}}=m_{1}p_{1}(1-p_{1})-ep_{1}}$

${\displaystyle {\frac {dp_{2}}{dt}}=m_{2}p_{2}(1-p_{1}-p_{2})-ep_{2}-mp_{1}p_{2}}$

In this case, if e is too high, p₁ and p₂ will be zero at steady state. However, when the rate of extinction is moderate, p₁ and p₂ can stably coexist. The steady state value of p₂ is given by

${\displaystyle p_{2}^{*}={\frac {e}{m_{1}}}-{\frac {m_{1}}{m_{2}}}}$

(p*₁ may be inferred by symmetry). If e is zero, the dynamics of the system favor the species that is better at colonizing (i.e. has the higher m value). This leads to a very important result in theoretical ecology known as the Intermediate Disturbance Hypothesis, where the biodiversity (the number of species that coexist in the population) is maximized when the disturbance (of which e is a proxy here) is not too high or too low, but at intermediate levels.^[44]

The form of the differential equations used in this simplistic modelling approach can be modified. For example:

1. Colonization may be dependent on p linearly (m*(1-p)) as opposed to the non-linear m*p*(1-p) regime described above. This mode of replication of a species is called the “rain of propagules”, where there is an abundance of new individuals entering the population at every generation. In such a scenario, the steady state where the population is zero is usually unstable.^[45]
2. Extinction may depend non-linearly on p (e*p*(1-p)) as opposed to the linear (e*p) regime described above. This is referred to as the “rescue effect” and it is again harder to drive a population extinct under this regime.^[45]

The model can also be extended to combinations of the four possible linear or non-linear dependencies of colonization and extinction on p are described in more detail in.^[46]

Food webs

Food webs provide a framework within which a complex network of predator–prey interactions can be organised. A food web model is a network of food chains. Each food chain starts with a primary producer or autotroph, an organism, such as a plant, which is able to manufacture its own food. Next in the chain is an organism that feeds on the primary producer, and the chain continues in this way as a string of successive predators. The organisms in each chain are grouped into trophic levels, based on how many links they are removed from the primary producers. The length of the chain, or trophic level, is a measure of the number of species encountered as energy or nutrients move from plants to top predators.^[48] Food energy flows from one organism to the next and to the next and so on, with some energy being lost at each level. At a given trophic level there may be one species or a group of species with the same predators and prey.^[49]

In 1927, Charles Elton published an influential synthesis on the use of food webs, which resulted in them becoming a central concept in ecology.^[50] In 1966, interest in food webs increased after Robert Paine's experimental and descriptive study of intertidal shores, suggesting that food web complexity was key to maintaining species diversity and ecological stability.^[51] Many theoretical ecologists, including Sir Robert May and Stuart Pimm, were prompted by this discovery and others to examine the mathematical properties of food webs. According to their analyses, complex food webs should be less stable than simple food webs.^{[1]: 75–77 [2]: 64} The apparent paradox between the complexity of food webs observed in nature and the mathematical fragility of food web models is currently an area of intensive study and debate. The paradox may be due partially to conceptual differences between persistence of a food web and equilibrial stability of a food web.^[1][2]

Systems ecology

Systems ecology can be seen as an application of general systems theory to ecology. It takes a holistic and interdisciplinary approach to the study of ecological systems, and particularly ecosystems. Systems ecology is especially concerned with the way the functioning of ecosystems can be influenced by human interventions. Like other fields in theoretical ecology, it uses and extends concepts from thermodynamics and develops other macroscopic descriptions of complex systems. It also takes account of the energy flows through the different trophic levels in the ecological networks. Systems ecology also considers the external influence of ecological economics, which usually is not otherwise considered in ecosystem ecology.^[52] For the most part, systems ecology is a subfield of ecosystem ecology.

Swarm behaviour

Swarm behaviour is a collective behaviour exhibited by animals of similar size which aggregate together, perhaps milling about the same spot or perhaps migrating in some direction. Swarm behaviour is commonly exhibited by insects, but it also occurs in the flocking of birds, the schooling of fish and the herd behaviour of quadrupeds. It is a complex emergent behaviour that occurs when individual agents follow simple behavioral rules.

Recently, a number of mathematical models have been discovered which explain many aspects of the emergent behaviour. Swarm algorithms follow a Lagrangian approach or an Eulerian approach.^[56] The Eulerian approach views the swarm as a field, working with the density of the swarm and deriving mean field properties. It is a hydrodynamic approach, and can be useful for modelling the overall dynamics of large swarms.^[57][58][59] However, most models work with the Lagrangian approach, which is an agent-based model following the individual agents (points or particles) that make up the swarm. Individual particle models can follow information on heading and spacing that is lost in the Eulerian approach.^[56][60] Examples include ant colony optimization, self-propelled particles and particle swarm optimization.

On cellular levels, individual organisms also demonstrated swarm behavior. Decentralized systems are where individuals act based on their own decisions without overarching guidance. Studies have shown that individual Trichoplax adhaerens behave like self-propelled particles (SPPs) and collectively display phase transition from ordered movement to disordered movements.^[61] Previously, it was thought that the surface-to-volume ratio was what limited the animal size in the evolutionary game. Considering the collective behaviour of the individuals, it was suggested that order is another limiting factor. Central nervous systems were indicated to be vital for large multicellular animals in the evolutionary pathway.

Synchronization

Photinus carolinus firefly will synchronize their shining frequencies in a collective setting. Individually, there are no apparent patterns for the flashing. In a group setting, periodicity emerges in the shining pattern.^[62] The coexistence of the synchronization and asynchronization in the flashings in the system composed of multiple fireflies could be characterized by the chimera states. Synchronization could spontaneously occur.^[63] The agent-based model has been useful in describing this unique phenomenon. The flashings of individual fireflies could be viewed as oscillators and the global coupling models were similar to the ones used in condensed matter physics.