Abstract
Systematic Conservation Planning (SCP) is a widely-used approach to develop networks of protected areas. A crucial step in the SCP process is to set conservation targets for biodiversity features (explicit goals that quantify the minimum amount of each biodiversity feature to be covered by the protected areas). When the biodiversity features are different habitats occurring in the planning region, a relevant approach, based on the Species-Area Relationship (SAR), defines targets so as to maximize biodiversity representation within each habitat type. While many formulations of the SAR exist, their application remains dominated by the log-transformation of Power-law model. However, documented habitat-related and taxonomic idiosyncrasies in the shape of the SAR question the effectiveness of a given ubiquitous model in fitting data compared to others. Here, using 13 SAR functional forms, we investigate whether the habitat-related SAR uncertainties propagate across the entire conservation planning process and lead to both divergent conservation targets and conservation solutions for six habitats in the Mediterranean sea. Results revealed uncertainties in model selection across habitats, which leads to different SAR habitat-targets. Constraining a systemic conservation planning tool (Marxan) with those targets provided contrasted sets of priority areas for different SAR scenario. Our study demonstrated that restraining to one particular SAR model is inappropriate at fitting all SAR datasets, providing consequently conservation targets diverging markedly from data-driven SAR inferences. More importantly, corresponding reserve networks are either inefficient or overstated for the protection of habitats, leading to waste of scarce conservation resources that should be used sparingly. Therefore, we suggest to evaluate different SAR models and, when appropriate to carry out a multi-model inference to provide robust habitat-specific conservation targets.