A probability-based approach to match species with reserves when data are at different resolutions

Gap analysis is a protocol for assessing the extent to which valued biodiversity attributes are represented within protected areas. Such analysis involves overlaying the distribution of biodiversity features (e.g. species) with protected areas, but the protocol entails arbitrary assumptions that affect the outcome of the assessments. In particular, since species' distributions are usually mapped at a coarser resolution than protected areas, rules have to be defined to match the two data layers. Typically, a grid cell is considered protected if a given proportion is covered by protected areas. Because the effectiveness of protected areas is dependent on the definition of such arbitrary proportions (i.e., thresholds), errors of commission and omission in the level of species' representation are bound to exist. We propose an alternative approach whereby the contribution of a cell for the representation of species is defined as the expected value of a hyper-geometric random variable. We compare the conventional approach based on fixed thresholds with this new probability-based approach for both static and dynamic conservation scenarios, using a virtual dataset and a 100-plant-species' dataset for Iberian Peninsula. Results support the view that traditional fixed thresholds yield inconsistent results. Because species present different distributional patterns coinciding differently with protected areas, species-specific and time-specific thresholds should be used. Our approach enables to easily obtain these more adequate threshold values, thus offering a promising method for gap analyses. Future studies should seek to evaluate the performance of this method empirically in different conservation planning contexts.