Abstract

Networks of no-take marine reserves, and other marine protected areas, are becoming important management tools as the scope of fisheries broadens toward an ecosystem-based approach and merges with the concerns and practices of coastal zone management. While the basic principles of reserve design are simple (representation, replication, connectivity at ecologically relevant scales), in practice the selection of areas for protection is complex due to the high number of ecological and sociological factors involved, the incomplete nature of most data sets, and potential for conflicting goals. Site selection, thus, involves a multivariable system were each element can be differently considered according to the local characteristics and/or needs. Numerical models can be used as a tool to realize such evaluations in an objective manner based on predetermined assumptions and goals. However, to meet basic design principles, model implementation requires that the available data (e.g., habitat distributions, bathymetry, etc.) and scale of analysis are structured so that the relevant ecology of the system is accounted for. We show how this approach can be implemented from existing data in the US Caribbean using MARXAN. We then apply the model to explore the affects of different goals and assumptions. Results show that given the available data in the appropriate format, MARXAN can identify areas maximizing biodiversity conservation at minimum cost while maintaining basic design principles. However, missing data, and the incorporation of specific species for conservation (e.g., manatees), variable conservation goals (e.g., fully protecting all known spawning aggregations) and variable costs (preferentially sitting reserves in existing MPAs) have significant impacts on results. This suggests that stakeholders using results from MARXAN, or other numerical models, should be fully cognizant of how the program functions and the basic goals and assumptions employed prior to interpreting results.

PDF Preview