IENE 2016

For their survival, animal species depend on networks of habitat patches between which they can easily move (i.e. habitat networks). Similarly, human well-being also in part depends on networks of settlements that are well-connected by roads and traffic (i.e. human networks). However, human networks often pose a threat to the connectivity in habitat networks. This threat is two-fold; animal habitats and settlements usually are mutually exclusive (i.e. settlements and natural habitats rarely occur at the same location) and in most cases traffic intensity negatively influences the movement of animals between habitats. Due to the interactions between these coupled spatial networks, it is difficult to determine how the configuration of the human network affects connectivity in the habitat networks. For instance, changes to the settlement configuration may not only directly influence habitat connectivity, but also indirectly via changes in traffic flows. Although such knowledge is important for sustainable landscape and transport planning, few studies have assessed how habitat connectivity is influenced by settlement and road network configuration in coupled human and habitat networks. In this study, we have developed a model with which habitat connectivity is simulated at a regional scale in landscapes with variable settlement and road network configurations. In binary landscapes, we independently varied the number of settlements, the size of the largest settlement patch as well as the total proportion of settlement. The settlements were embedded in a matrix of continuous habitat. Settlements were connected with either dense or sparse road networks. On each road, expected traffic volumes were calculated with novel radiation models. Similar to the road network, the habitat network was constructed by connecting neighbouring habitat cells. With an animal dispersal model, we calculated the probability that an organism would successfully move between cells, which depended on the distance and traffic volumes between cells. Such habitat networks were parameterised for a range of species with different dispersal abilities and sensitivity to traffic (i.e. tree frog, hedgehog and badger). With this setup, the topology of a habitat network and the strengths of its links were influenced by settlement configuration and traffic volumes, respectively. From the habitat networks, we calculated overall habitat connectivity with two commonly used connectivity measures. In general, we found a negative correlation between the number of settlements and the habitat connectivity. Furthermore, for landscapes with a high proportion of settlement, it was more beneficial for habitat connectivity to have one very large settlement patch surrounded by smaller ones. Surprisingly, for some simulations, we found that habitat connectivity was higher with dense road networks than with sparse networks. In addition to new insights in the ecological processes governing habitat connectivity, the results from this study can provide valuable information for the planning of new settlement and transport networks as well as for the improvement of existing networks. Whereas the current study was based on a simplified simulation model, in a follow-up project we aim to add more realism to our models by coupling empirically derived habitat connectivity networks with land-use transport interaction models.