Backup coverage models in nature reserve site selection with spatial spread risk heterogeneity

Set covering and maximal covering problems are well-known integer programming models in location analysis. Such models have also been used in reserve site selection modeling. They aim at selecting sites to conserve species, sometimes reflecting a desire to group protected sites together or to separate sites. This paper uses such models but considers the case of land heterogeneity in terms of the risk of large disturbances that threaten species even within a reserve, such as fires, diseases, pests or invasive species. It removes the classical assumption of homogeneous land sites and considers both adjacency in areas with a low risk of multi-parcel disturbance and distance between sites in areas with a high risk of a large destructive event. The models are explored in a stylized data set and applied to a portion of the state of Oregon with comparison between the standard covering models in homogeneous and heterogeneous risk settings.