Spatially dynamic species are species whose populations occupy different areas at different times. Their long term survival often depends on the connections between these areas. Conservation strategies for these species therefore require realistic models of these connections. Furthermore, conservation strategies must be designed at a sufficiently large scale that we can make efficient decisions for a species throughout its range. However, models of spatial dynamics are rarely, applied at these scales. We present two new approaches for integrating spatial population dynamics into large-scale conservation planning.

Firstly, spatial dynamic models require high resolution data on species and habitat distributions. Distribution modeling can help provide this but may not be robust for species with a high rate of ‘false absences’ (i.e. locations which are suitable for a species but from which it is not recorded). This is an inherent problem when modeling species whose spatial dynamics cause them to exists in metapopulations. These species do not occupy all suitable habitat at all times. We addressed this problem for the Marsh Fritillary butterfly, Euphydryas aurinia, in Wales, UK. We designed high resolution habitat models based on the distribution of suitable habitat, rather than the species itself. Secondly, we used these distribution models in conjunction with a stochastic patch occupancy metapopulation model to investigate the effects of different reserve designs.
Results/Conclusions

Habitat-based distribution models had significantly greater discrimination power (mean AUC = 0.84) than species-based models (mean AUC = 0.69), and were better able to generalize to novel areas and large spatial extents. The current level of habitat protection is insufficient to ensure the long-term survival of the Marsh Fritillary. The most effective reserve design was a novel one which concentrates resources on a few areas but facilitates connections between very distant populations. Thus, it buffers effectively against stochastic extinction events. Simulations show that this solution considerably improves the amount of habitat occupied compared to other solutions, and gives a high level of confidence in the long term survival of the species.