Spatial zoning design for marine protected areas through multi-objective decision-making

• We develop two Multi-Objective Integer Linear Programming (MOILP) models for MPA zoning.
• MOILP guarantee to find global optima of MPA spatial zoning among various objects.
• Each model has its own characteristics to fit in special MPA design preferences.
• Our models have certain advantages over the popular Marxan with zones software.
• We provide good alternatives for MPA spatial zoning by using the proposed MOILP models.

Systematic approaches to designing marine protected areas (MPA) have recently garnered more attention as an efficient planning tool. In particular, spatial zoning for various types of MPA designs could accommodate several demands for ecological conservation while minimizing socioeconomic impact. Most spatial zoning approaches are formulated as nonlinear Multi-Objective Decision-Making (MODM) models, which are solved using stochastic search algorithms, such as the popular Marxan with Zones. Due to the stochastic nature of these algorithms, the final MPA design is the composite result of several modeling experiments, and the solution is often suboptimal. For the current study, two MODM models were developed based on Multi-Objective Integer Linear Programming (MOILP) for MPA spatial zoning. The proposed models are referred to as a buffer cells model (BCM) and an external border cells model (EBCM). BCM allocates two types of cells to an MPA zone, covering core and buffer cells. The EBCM uses external border cells instead of buffer cells. Both models can minimize the cost incurred by MPA implementation while simultaneously satisfying different conservation targets. The solutions found are globally optimal. MPA designs from the BCM, EBCM, and Marxan with Zones are compared by displaying their spatial distributions. The results present the general characteristics of the BCM and EBCM and demonstrate how both models may have certain advantages over the Marxan with Zones method and can thus be considered as good alternatives for MPA spatial zoning.