Multi-Objective Methods for Reserve Site Selection: A Case Study Selecting Priority Sites for Conservation in Belize

Abstract

Systematic conservation planning encompasses a range of data-driven optimisation techniques that are often used to prioritise geographical areas in terms of ecological significance. This evaluation is usually based on criteria that represent the socioeconomic costs associated with conserving specific sites, the landscape properties of the proposed reserve networks and various measures for the included biodiversity features. In this study, biological and socioeconomic data for the lowland savannas of Belize have been combined into layers that reflect the three aforementioned properties and the reserve design problem has been described in a purely multi-objective way that does not require the designation of arbitrary weights for the various criteria. Subsequently, the applicability of two multi-objective optimisation algorithms in solving this version of the reserve design problem for Belizean lowland savannas is explored. It is hypothesized that the multi-objective methods may be able to indicate reserve network configurations that are representative of the full range of potential trade-offs between selected objectives. The results of the implemented algorithms are compared to a series of solutions obtained by MARXAN, a well-established conservation planning software tool. The findings are discussed in detail and indicate that the two methods are effective in approximating certain areas of the trade-off surface between different objectives, but do not produce solutions across its entirety according to the original hypothesis. Two factors that contribute to loss of solution diversity are identified and attributed to the effect of certain implementation choices on the evolvability of the optimisation algorithms.