Niche shift versus niche conservatism? Climatic characteristics of the native and invasive ranges of the Mediterranean house gecko (<i>Hemidactylus turcicus</i>)

ABSTRACT Aim The use of ecological niche models (ENMs) to predict potential distributions of species is steadily increasing. A necessary assumption is that climatic niches are conservative, but recent findings of niche shifts during biological invasion indicate that this assumption is not always valid. Selection of predictor variables may be one reason for the observed shifts. In this paper we assess differences in climatic niches in the native and invaded ranges of the Mediterranean house gecko ( Hemidactylus turcicus ) in terms of commonly applied climate variables in ENMs. We analyse which variables are more conserved versus relaxed (i.e. subject to niche shift). Furthermore, we study the predictive power of different sets of climate variables. Location The Mediterranean region and North America. Methods We developed models using Maxent and various subsets of variables out of 19 bioclimatic layers including: (1) two subsets comprising almost all variables excluding only highly collinear ones; (2) two subsets with minimalistic variable sets of water availability and energy measures; (3) two subsets focused on temperature‐related parameters; (4) two subsets with precipitation‐related parameters; and (5) one subset comprising variables combining temperature and precipitation characteristics. Occurrence data from the native Mediterranean range were used to predict the potential introduced range in North America and vice versa. Degrees of niche similarity and conservatism were assessed using both Schoener's index and Hellinger distances. The significance of the results was tested using null models. Results The degree of niche similarity and conservatism varied greatly among the predictors and variable sets applied. Shifts observed in some variables could be attributed to active habitat selection while others apparently reflected background effects. Main conclusions The study was based on comprehensive occurrence data from all regions where Hemidactylus turcicus is present in Europe and North America, providing a robust foundation. Our results clearly indicate that the degree of conservatism of niches in H. turcicus largely varies among predictors and variable sets applied. Therefore, the extent of niche conservatism of variables applied should always be tested in ENMs. This has an important impact on studies of biological invasion, impacts of climate change and niche evolution.