Do species population parameters and landscape characteristics affect the relationship between local population abundance and surrounding habitat amount?

摘要

In landscape ecology, correlational approaches are typically used to analyse links between local population abundance, and the surrounding habitat amount to estimate biologically-relevant landscape size (extent) for managing endangered or pest populations. The direction, strength, and spatial extent of the correlations are then sometimes interpreted in terms of species population parameters. Here we simulated the population dynamics of generalized species across spatially explicit landscapes that included two distinct habitat types. We investigated how characteristics of a landscape (structure), including the variation in habitat quality and spatial aggregation of the habitat, and the precise population-dynamic properties of the simulated species (dispersal and growth rates) affect the correlation between population abundance and amount of surrounding favourable habitat in the landscape. To evaluate these spatial extents of correlation, proportions of favourable habitat were calculated within several circles of increasing diameter centred on sampling patches of favourable habitat where population abundance was recorded. We found that the value of the correlation coefficients between population abundance and amount of surrounding favourable habitat depended on both population dynamic parameters and landscape characteristics. Coefficients of correlation increased with the variation in habitat quality and the aggregation of favourable habitat in the landscape, but decreased with the dispersal distance. The distance at which the correlation was maximized was sensitive to an interaction between the level of aggregation of the habitat and the dispersal distance; whereas the greatest distance at which a significant correlation occurred was more sensitive to the variation in habitat quality. Our results corroborate the view that correlational analyses do provide information on the local population dynamics of a species in a given habitat type and on its dispersal rate parameters. However, even in simplified, model frameworks, direct relationships are often difficult to disentangle and global landscape characteristics should be reported in any studies intended to derive population-dynamic parameters from correlations. Where possible, replicated landscapes should be examined in order to control for the interaction between population dynamics and landscape structure. Finally, we recommend using species-specific, population-dynamic modelling in order to interpret correctly the observed patterns of correlation in the landscape.