Evidence of climatic niche shift during biological invasion O. Broennimann, U. A. Treier, H. Müller-Schärer, W. Thuiller, A. T. Peterson, and A. Guisan Abstract: Niche-based models calibrated in the native range by relating species observations to climatic variables are commonly used to predict the potential spatial extent of species' invasion. This climate matching approach relies on the assumption that invasive species conserve their climatic niche in the invaded ranges. We test this assumption by analysing the climatic niche spaces of Spotted Knapweed in western North America and Europe. We show with robust cross-continental data that a shift of the observed climatic niche occurred between native and non-native ranges, providing the first empirical evidence that an invasive species can occupy climatically distinct niche spaces following its introduction into a new area. The models fail to predict the current invaded distribution, but correctly predict areas of introduction. Climate matching is thus a useful approach to identify areas at risk of introduction and establishment of newly or not-yet-introduced neophytes, but may not predict the full extent of invasions. The study examined the climatic niche of Centaurea maculosa in its native and invaded ranges to test whether the species exhibits niche conservatism, a pivotal assumption for enabling reciprocal geographic predictability between the two ranges. Comprehensive occurrence data from all regions where the species is present in Europe and western North America were used, fully covering the relevant large climatic gradients and eliminating risk of fitting truncated response curves and thus only partially fitting models to the species' realized niche. The niche defined here reflects all climatic conditions where the plant can survive and reproduce in the presence of biotic interactions. We are not aware of comparably robust data for any other invasive species across two ranges. The analyses provided the same results and supported the same conclusions whatever the climatic data set used. Only the results conducted with the eight CRU 10' climatic maps are presented here, because these were considered biologically more relevant for the species and they constitute the baseline data set used by the Intergovernmental Panel for Climate Change and were already used in similar studies. The eight CRU 10' climatic maps were: ratio of actual to potential evapotranspiration (aet/pet), potential evapotranspiration (pet), annual amount of precipitations (prec), annual variation of precipitations (std_prec), minimum temperature of the coldest month (tmin), annual mean temperature (tmp), maximum temperature of the warmest month (tmax) and growing degree-days above 5 °C (gdd). Results obtained with other climatic data sets are available in the Supplementary Material. Principal component analysis of the pooled climatic data revealed two significant axes of climatic variation, defining a realized climate space of reduced dimensionality which allows the investigation of niche conservatism. The enclosed correlation circle indicates the relativeEvidence of climatic niche shift during biological invasion O. Broennimann, U. A. Treier, H. Müller-Schärer, W. Thuiller, A. T. Peterson, and A. Guisan Abstract: Niche-based models calibrated in the native range by relating species observations to climatic variables are commonly used to predict the potential spatial extent of species' invasion. This climate matching approach relies on the assumption that invasive species conserve their climatic niche in the invaded ranges. We test this assumption by analysing the climatic niche spaces of Spotted Knapweed in western North America and Europe. We show with robust cross-continental data that a shift of the observed climatic niche occurred between native and non-native ranges, providing the first empirical evidence that an invasive species can occupy climatically distinct niche spaces following its introduction into a new area. The models fail to predict the current invaded distribution, but correctly predict areas of introduction. Climate matching is thus a useful approach to identify areas at risk of introduction and establishment of newly or not-yet-introduced neophytes, but may not predict the full extent of invasions. The study examined the climatic niche of Centaurea maculosa in its native and invaded ranges to test whether the species exhibits niche conservatism, a pivotal assumption for enabling reciprocal geographic predictability between the two ranges. Comprehensive occurrence data from all regions where the species is present in Europe and western North America were used, fully covering the relevant large climatic gradients and eliminating risk of fitting truncated response curves and thus only partially fitting models to the species' realized niche. The niche defined here reflects all climatic conditions where the plant can survive and reproduce in the presence of biotic interactions. We are not aware of comparably robust data for any other invasive species across two ranges. The analyses provided the same results and supported the same conclusions whatever the climatic data set used. Only the results conducted with the eight CRU 10' climatic maps are presented here, because these were considered biologically more relevant for the species and they constitute the baseline data set used by the Intergovernmental Panel for Climate Change and were already used in similar studies. The eight CRU 10' climatic maps were: ratio of actual to potential evapotranspiration (aet/pet), potential evapotranspiration (pet), annual amount of precipitations (prec), annual variation of precipitations (std_prec), minimum temperature of the coldest month (tmin), annual mean temperature (tmp), maximum temperature of the warmest month (tmax) and growing degree-days above 5 °C (gdd). Results obtained with other climatic data sets are available in the Supplementary Material. Principal component analysis of the pooled climatic data revealed two significant axes of climatic variation, defining a realized climate space of reduced dimensionality which allows the investigation of niche conservatism. The enclosed correlation circle indicates the relative