Abstract
Habitat suitability estimated with probability of occurrence in species distribution models (SDMs) is used in conservation to identify geographic areas that are most likely to harbor individuals of interest. In theory, probability of occurrence is coupled with individual fitness so that individuals have higher fitness at the centre of their species environmental niche than at the edges, which we here define as ‘fitness-centre' hypothesis. However, such relationship is uncertain and has been rarely tested across multiple species. Here, we quantified the relationship between coarse-scale probability of occurrence projected with SDMs and individual fitness in 66 tree species native of North America. We used 1) field data of individuals' growth rate (height and diameter standardized by age) available from the United States Forest Inventory Analysis plots; and 2) common garden data collected from 23 studies reporting individual growth rate, survival, height and diameter of individuals originated from different provenances in United States and Canada. We show ‘fitness–centre' relationships are rare, with only 12% and 11% of cases showing a significant positive correlation for field and common garden data, respectively. Furthermore, we found the ‘fitness–centre' relationship is not affected by the precision of the SDMs and it does not depend upon dispersal ability and climatic breath of the species. Thus, although the ‘fitness–centre' relationship is supported by theory, it does not hold true in nearly any species. Because individual fitness plays a relevant role in buffering local extinction and range contraction following climatic changes and biotic invasions, our results encourage conservationists not to assume the ‘fitness–centre' relationship when modelling species distribution.