A history of metapopulation competition models
Space was already regarded as an important determinant of competitive interactions early in the modern history of competition theory. The works by Levins (1969,1970), Levins and Culver (1971), Horn and MacArthur (1972), and Levin (1974) were particularly influential.
Most of these focused on multi-patch systems and ignored the details of competition within a patch; competition was simply assumed to increase the extinction rate of each consumer species by a fixed amount in patches that were occupied by both. Levin (1974) had a treatment of 2-patch systems that included within-patch dynamics rather than just presence/absence. He called attention to the fact that the alternative-exclusion outcome of the LV model allowed two competitors to coexist in a landscape in which the different histories of the two patches led to a different species being dominant in each one. Most experimental approaches to competition did not explore such subdivided systems until many years later.Tilman (1982) stressed that different spatial conditions could allow any number of competitors to coexist, although he mainly considered the special case in which the dynamics of two limiting resources varied spatially and he did not include any explicit models of movement between patches. A variety of spatial models of plant competition were developed in the 1990s (reviewed in Pacala and Levin 1997, and other chapters in Tilman and Kareiva 1997). Many of these used the individual plant as the spatial unit, so were dealing with systems in which space itself was the resource. While these works produced some important insights, they had relatively little influence on approaches to competition in systems other than terrestrial plants.
Movement has long been recognized as having some counterintuitive effects on total abundance in a metapopulation. The first study to examine this possibility was Holt (1985).
He showed that random movement by consumers between two unequal patches altered intraspecific competition and had counter-intuitive effects on equilibrium consumer abundance. The evolutionarily stable movement rate in a constant environment was shown to be zero. Movement rates greater than zero were not favoured at the level of individual selection. However, in the simple constant-environment consumer-resource models that Holt explored, the presence of some consumer movement often reduced overexploitation of the more productive patch(es), and thereby increased the global equilibrium population size.There were relatively few studies of different types of movement on interspecific interactions until the concept of metacommunities became popular, shortly before 2000. The most common approach at that time was to assume a constant per capita movement rate of individuals out of each patch. Their destination in multi-patch systems was usually assumed to be equally likely to be any other patch, although some works assumed higher probabilities of arriving at closer patches. Subsequent work devoted more attention to the possibility of adaptive movement (Abrams 2000b, 2007a; Amarasekare 2010). As Ronce (2007, p. 233) put it, ‘There are good theoretical reasons to believe that informed dispersal decisions would confer an evolutionary advantage over a blind process, unless patterns of variation in habitat quality are totally unpredictable or information acquisition is costly’. The process of developing theory that reflects this generalization has been complicated by the many possible ways of modelling adaptive movement (discussed in Abrams et al. (2007), Amarasekare (2010), and Gross et al. (2020), among others).
10.3