Technical Comments

Response to Comment on “Climatic Niche Shifts Are Rare Among Terrestrial Plant Invaders”

Science  12 Oct 2012:
Vol. 338, Issue 6104, pp. 193
DOI: 10.1126/science.1226051

Abstract

Webber et al. take a critical view of our findings that niche expansions are rare in plant invaders, arguing mainly that we did not include nonanalog climates in our analyses. Yet, their concerns include misunderstandings and go beyond the scope of our study, which was purposely restricted to analog climates. We further explain why our results remain robust to other factors of niche dynamics in the native range. We conclude that the implications of our findings remain valid for projections of niche models in analog climates and that projections in nonanalog climates should be undertaken with care.

Webber et al. (1) take a critical view of our findings that niche expansions are rare in plant invaders when only climates available in both native and invaded ranges are considered (2). Their main point of contention is that we did not include in our analyses those climates in the invaded range that are not available in the native range (i.e., nonanalog climates). Webber et al. argue that niche shifts could occur in nonanalog climates and that they are important when making predictions with niche models. They also argue that other factors in the native range could explain these niche changes. However, their comments mix different issues, misinterpreting our findings and going far beyond the scope that we defined for our results and conclusions. We reply here to their main points.

First, we never concluded that a niche expansion—especially of the realized niche as Webber et al. argue—could not occur in nonanalog climates. Indeed, as we reported, our niche change quantification framework allows all situations to be quantified, including both analog and nonanalog situations, and we reported [(2) and its supporting online material] that niche analyses in total climate reveal more shift. We chose, however, not to report on analyses involving nonanalog situations in the main report because their interpretation remains highly speculative (35). There, only complementary experimental approaches could help determine whether the expansion is caused by a real change of the fundamental niche or is due to preadaptation (6, 7), because the fundamental niche cannot be quantified using empirical distribution data (8). Indeed, when Webber et al. state, “Such expansion can arise when the new location has nonanalogous climates but where the values of the key climatic variables…still fall within the species’ physiological limits,” they give credit to our approach of not considering nonanalog environments, because what they describe is precisely the preadaptation situation just discussed. In this regard, their speculative case of a Mediterranean species invading tropical areas would also be an example of preadaptation. These nonanalog situations can indeed be quantified with our framework, but they cannot be ascertained as being an expansion of either the fundamental or realized niches because, without experiment, one cannot know whether the species would (e.g., preadaptation) or would not (e.g., outside the niche, dispersal limitations, or biotic exclusion) have occupied these environments if they had been available in the native range. Still, we acknowledge that these situations of niche filling in nonnative conditions are important when making spatial predictions and can be accommodated by fitting models with data from both native and invaded ranges (9, 10).

Second, we agree that dispersal limitation or biotic interactions can effectively result in geographic unfilling in the native range (11), but unfilling in geographic space does not translate into unfilling in climatic niche space. Furthermore, if native niche unfilling was occurring, we should indeed observe more niche expansion in the introduced range. More important, though, Webber et al. misunderstand our analyses and graphs when they write: “Furthermore, elements of infilling (star, Fig. 1), as quantified by Petitpierre et al. in their figure 3, are arguably uninformative to understanding the potential niche of the species,” because they compare two very different things. Whereas their figure 1 conceptually compares the niche of one species between ranges, our figure 3 is based on observations and represents the accumulation of events of (i) niche expansion versus stability and (ii) niche unfilling versus stability across all species, as a way of summarizing the extent of these phenomena in climatic space. Therefore, holes in the niche of a single species (their star) simply cannot be observed in our graphs because there is no way to represent the niches of individual species there. Another complication of their figure 1 is that the hole they represented in the two-dimensional (2D) niche (their star) is not represented in the corresponding 1D response curves (in the native range), resulting in a discrepancy between their interpretation of niche changes along the principal component analysis axes (their gradient from red to yellow to blue) and the real niche in 2D. Furthermore, in our work we deliberately did not use the term “infilling” because the situation they describe (star in their figure 1) is unlikely to persist in our quantitative framework (12), which uses a kernel density smoother that aims precisely at avoiding these ecologically unrealistic situations of “holes” within a species’ niche. This framework builds on previous work [e.g., (13)] and therefore represents the latest advance in multivariate niche quantification, thus accounting for a modern, probabilistic definition of the niche, including older concepts such as the law of tolerance.

Third, we agree that studies of nonanalog climates would be useful to forecast effects of climate change and biological invasions, but these cannot rely on current distribution data [(13) and below]. As developed above, we avoided nonanalog situations because these were recently shown (3, 4, 14) to be conditions where predictions by ecological niche models should not be made [see (14) and the “C” area in its figure 1]. Following these recommendations, and contrary to claims by Webber et al., what our study truly demonstrated is that assessing climate change impact within the range of analogous climates is likely to be a valuable use of climate niche models. It is therefore curious to see the same reference (14) supporting our choice to exclude nonanalogous climates, also used by Webber et al. to question our conclusions.

To conclude, Webber et al. provide themselves the main conclusion to our reply when they write “their limited concept of expansion does not allow them to exhaustively explore whether species niches appear to shift or change in nonanalog environments.” We can only agree, as our study never pretended to address niche shifts in nonanalog situations. We further agree that complementary studies of niche changes in nonanalog situations would be particularly useful to assess the ability of niche models to project into nonanalog climates (in the future or in different ranges) and that these studies will need to rely on more mechanistic approaches based on experimental measurements, not only on empirical distribution data.

References and Notes

  1. Nonanalog environments are called nonoverlapping background by Peterson (3).
  2. Preadapation are those cases where the fundamental niche of the species would include conditions outside the available environment in the native range and thus would ensure adaptation to these conditions if present in the invaded range.
  3. Acknowledgments: A.G., B.P., and O.B. thank the Swiss National Centre for Competence in Research (NCCR) “Plant Survival in Agro-ecological Landscapes” for financial support.
View Abstract

Subjects

Navigate This Article