Modelling the loss of genetic diversity in vole populations in a spatially and temporally varying environment

Altering environmental conditions affects the genetic composition of populations via demographic and selective responses by creating of variety of population substructuring types. Classical genetic approaches can predict the genetic composition of populations under long-term or structurally stable conditions, but exclude factors such as animal behaviour, environmental structure, and breeding biology, all of which influence genetic diversity. Most populations are unique in some of these characteristics, and therefore may be unsuitable for the classical approach. Here, an alternative approach using a genetically explicit individual-based model (IBM) coupled to a dynamic landscape model was used to obtain measures for the genetic status of simulated vole populations. The rate of loss of expected heterozygosity (H-e) was calculated for simulated populations using two levels of spatial and temporal heterogeneity. Results showed that both spatial and temporal heterogeneity exerted an influence on the rate of loss of genetic diversity, but the precise effect was a balance between the effects of population sub-structuring, the frequency of founder effects and population size. These were in turn related to habitat availability and their influence on vole behaviour. Interaction between spatial and temporal dynamics altered the ratio of effective population size to census size. This indicates an altered reproductive potential, crucial in conservation biology applications. However, when the loss of heterozygosity was corrected for the harmonic mean of the population size, the rate of loss was almost identical in the four scenarios. Unlike classical genetic models, IBMs are flexible enough to mimic real population processes under a range of environmental and behavioural conditions. We conclude that IBMs incorporating explicit genetics provide a promising new approach to the evaluation of the effect of animal behaviour, and random and man-induced events on the genetic composition of populations. They also provide a new platform from which to investigate the implication of real world deviations from assumptions of traditional genetic models.