Refugia and dispersal promote population persistence under variable arid conditions: a spatio‐temporal simulation model

In arid environments, population dynamics of many organisms follow resource pulses in time and space. This heterogeneity in resource accessibility is due to irregular and local rainfalls. In drought periods, the population density of small mammals such as rodents falls so that animals seem absent across the landscape. How can they avoid extinction and persist in time and space at low densities during droughts? We hypothesize that a fraction of the population may survive in refugia—less arid patches—and recolonize the landscape after drought‐breaking rains. When precipitation and resources abound again, rodents become abundant and disperse over large areas. Our spatio‐temporal simulation tests the hypotheses that refugia and dispersal promote population persistence over large temporal and spatial scales. We programmed a virtual desert (100 × 100 matrix) in which a virtual population changes over the course of 100 time steps representing 100 years. In our simulations, when rainfall is scarce, refugia and dispersal are insufficient to produce population persistence, and when rainfall is heavy or widespread, these factors are not necessary. At moderate rainfall frequency, refugia and dispersal are essential for persistence, and long‐distance dispersers need fewer refugia than short‐distance ones. With cyclic rainfall patterns mimicking La Niña's influence on desert precipitation, long drought periods and short wet periods, persistence requires both abundant refugia and long distance dispersal.