干旱区生态环境知识资源中心

Aim The first aim of this paper was to evaluate the distribution of the three Sterocorax species found in the Iberian Peninsula by estimating the main environmental factors that constrain their distributions. The second aim was to explore the potential importance of competitive interactions in limiting their current distributions using predictive distribution models.

Location Iberian Peninsula.

Methods Species presence data were collected from records in the literature and private and public collections. Ecological niche factor analysis was performed to extract pseudo-absences (probable absences), which, together with presence data, were modelled using generalized additive models. The models were run twice. Initially we used only environmental variables, and thereafter additional spatial variables were included in order to account for spatially structured factors not accounted for in the environmental variables.

Results Highly reliable distribution models were obtained for the three species, with AUC scores (area under the receiver operating characteristics curve) higher than 0.96. The addition of spatial variables to the first model significantly improved the predicted distribution of Corax (Sterocorax) globosus and Corax (Sterocorax) insidiator, by reducing their potential distribution area. In contrast, the model of Corax (Sterocorax) galicianus was not improved by the addition of a spatial term.

Main conclusions Generated pseudo - absences, such as those used in this study, helped to avoid problems of using erroneous data (false absences) in distribution records. Pseudo-absences greatly improved the models by only selecting absences within the area with the most unfavourable environmental conditions. The importance of spatial variables to both C. (S.) globosus and C. (S.) insidiator distributions probably relates to a number of unknown factors, such as unique historical events. The absence of established populations of C. (S.) globosus north of the Ebro Valley appears to be one such historical factor. The distribution of C. (S.) galicianus only marginally overlaps with that of C. (S.) globosus, according to our environmental factor models. As this overlap is restricted it is not likely to be a result of competitive exclusion; rather, their geographical segregation seems to be environmentally mediated. The addition of spatial variables reduced the potential habitat of C. (S.) insidiator, eliminating some environmentally optimal areas from its distribution. As no environmental barrier seems apparent in this case, competitive interaction with C. (S.) globosus is a plausible hypothesis for its absence in these optimal parts of its range.