Alternative predictions for optimal dispersal in response to local catastrophic mortality (Q1897488)

Individual variation in dispersal ability has been described for a vast number of organisms, but it is the insects which provide us with our best studied examples; the expression of aptery (the flightless condition) has been shown to be affected by both habitat stability and density-dependent competition in a variety of wing dimorphic species. The importance of these two factors on observed dispersal patterns suggests that it may be appropriate to use ESS (evolutionarily stable strategy) models to predict optimal dispersal strategies. Not all organisms experience limited resources in nature and dispersal can have other advantages that are unrelated to the intensity of competition. For example, in highly disturbed habitats, population sizes may be kept quite low by catastrophic events and dispersal can represent an escape from nearly certain mortality at a local site. \textit{H. N. Comins} et al. [J. Theor. Biol. 82, 205-230 (1980)] generated the ESS prediction that the optimal proportion of dispersed offspring should vary directly with the probability that exogenous mortality would cause local extinction of all individuals at a site. They presented the explicit relationship between the dispersal fraction \(\nu_{\text{IND}}\) and the probability of local extinction \(X\) as \[ \nu_{\text{IND}} = X/ \bigl( 1 - (1 - X) p \bigr), \] where \(p\) is the probability of surviving dispersal. Although Comins et al. acknowledged the inadequacy of this overly simplistic prediction, they stated that ``the evaluation of \(\nu_{\text{IND}}\) provides a true lower limit to which the ESS rate must converge as (the number of adults per site) \(k \to \infty\).'' Hence, if there is no limit on the number of adults per site, competition should be weak and \(\nu_{\text{IND}}\) should provide an approximation of the optimal, density-independent response to catastrophic mortality. In contrast, using a branching process model, \textit{R. H. Karlson} and \textit{H. M. Taylor} [Theor. Popul. Biol. 42, No. 2, 218-233 (1992; Zbl 0757.92014)] predicted that the optimal proportion of dispersed offspring should vary inversely with the probability of local extinction. When local survivorship was quite low (i.e., near the limit where each parent was exactly replacing itself), the optimal dispersal strategy was to disperse no off-spring as opposed to the high proportion of dispersed offspring predicted by the ESS model. In order to understand more clearly this inconsistency in the predictions of the two models, we modified our original model to conform more closely with the assumptions of Comins et al. and then we generated a set of dispersal predictions to compare with their results. Below, we contrast the results of these models and suggest how to resolve the apparent conflict.