Reserve selection in regions with poor biological data

New approaches to the identification of priority areas for conservation are gaining popularity for their efficiency in maximizing species representation. However, their dependence on detailed distributional data severely hinders their application to regions where such information is limited, although these are commonly also the regions where conservation planning and action are most urgently required. We used exemplar data on the distribution of southern African birds to assess how sampling effort affects the performance of reserve networks selected by methods based on complementarity. We derived four scenarios of data availability from the initial data, resulting from different levels of sampling effort: abundance data, presence/absence data, low sampling effort, and absence of data. Reserve selection based on data obtained with low sampling effort can be highly effective in the representation of species, with a good relative performance also in terms of representation of species in peaks of abundance. This is because although the data on low sampling effort represent far fewer records than the original data, the records retained are biased toward the selection of peaks of abundance, even for the restricted-range species. Although the best results were naturally obtained from the most effort-intensive data set (with abundance data), these results suggest that methods based on complementarity are potentially valuable tools for reserve selection in regions for which biological data are poor.