Species distributions do not reflect climatic tolerance

Background/Question/Methods

The distribution of suitable habitat for many species has been projected to shift rapidly with climate change, likely causing local extirpations and ultimately widespread extinctions of species.  But, many of these projections rely on the assumption that regional species distributions are limited solely by climate.  Yet, several recent studies suggest that non-climatic factors, such as dispersal limitation, play a more important role in defining species ranges than previously thought.  Here, we tested whether occurrences in the native range effectively circumscribe the climatic conditions that species can survive.  We identified a set of 144 plants endemic to the US, but occurring outside of their native ranges due to introductions via the horticultural trade.  We used geographic occurrences from national herbaria, and identified the native range based on the Biota of North America Program (BONAP), which designates native ranges at the county level.  We then compared annual precipitation, minimum and maximum temperature between occurrences in the native and adventive ranges (occurrences outside the native range) to test whether the native range is a good indicator of climatic tolerance.  Finally, we modeled each species distribution using Maxent to quantify how much the addition of climate conditions found in the adventive range altered projections of suitable habitat.

Results/Conclusions

Distributions of species in their native ranges do not reflect their climatic tolerance.  For the median species, climate conditions at adventive occurrences had minimum temperatures 2.9 ˚C colder and maximum temperatures 1.2 ˚C warmer than native occurrences.  Adventive occurrences also appeared to tolerate much lower water availability, with the median species found in conditions 23 cm of annual precipitation drier than expected based on native range occurrences.  These broader climate limits translated into a median expansion of potential geographic range by 35% based on species distribution models.  Our results suggest that plants’ native ranges underestimate climatic tolerance, leading species distribution models to underpredict potential range.  Most species could likely occur much more broadly if climate were the only factor limiting their distributions.  These results suggest that some plants might be able to persist in situ with climate change for longer than projected by ecological forecasting models.  However, given this strong evidence of non-climatic range limits, it appears unlikely that distributions will be able to ‘keep up’ with climate change in the long-term, making narrowly distributed and dispersal limited species particularly vulnerable.