Systematic landscape restoration using integer programming

In some agricultural regions, reservation of remnant natural lands will not maintain the natural biodiversity and large-scale ecological restoration is required. Geographic planning for restoration is essential to obtain the maximum ecological benefit from the limited resources available for landscape restoration. In this paper, we present a proof-of-concept that implements principles of systematic conservation planning, such as adequacy, representativeness, efficiency and flexibility, within an integer programming framework to identify geographic priorities for landscape restoration in a small catchment in South Australia. Integer programming ensures maximally efficient solutions that meet conservation targets in the form of minimum proportions (30%) and minimum areas (15 ha) of each environment type. However, the optimum solution identifies a system of scattered sites with an inadequate and impractical landscape structure. We propose a system of impedance surfaces that prioritise the spatial location of restoration towards improving landscape structure. Different options for systematic landscape restoration are assessed and the costs of spatial prioritisation of areas is found to be either optimal or very near optimal regardless of the impedances used. Thus, the cost of spatial prioritisation of sites for restoration in the landscape is negligible. We suggest the proposed methodology fits within a broader plan for landscape restoration and natural resource management.