Abstract

Numerical models are tools used to identify areas of complex biodiversity or potential hotspots of fisheries production that then can be targeted for priority protection. On a larger scale these can be linked to form potentially self-sustaining habitat networks. Traditionally, models have used habitat as a surrogate for species or community representation, but have not addressed the more difficult task of ensuring that ecological function is incorporated into model results. We have identified an approach to structuring habitat data that facilitates the incorporation of ecological function into model outputs, as well as developing connectivity-based guidelines for assessing results. These were applied to data from Puerto Rico using Marxan. Model runs were made under two levels of clustering, with the “conservation target” arbitrarily set at 30 %. Results showed that only with higher clustering did priority areas meet the connectivity criteria, but at the cost of requiring about 50% more area to be selected. To further assess results, we constructed a “null” model composed of the four basic habitats (reef, sand, SAV, mangrove), which assumes that all patches within habitat type are equal. Results show little correlation between priority areas chosen by the two models, and patterns of frequency count, indicated that significant adjustments in area selection were made to incorporate ecological function. Again, additional costs were evident. Com-pared to the null model, the resulting number of planning units selected under the ecological function approach increased by 30%, regardless of the degree of clustering. The benefits are worth such costs.

PDF Preview