Print PageThere is an urgent need to predict how species will shift their geographic ranges in response to climate change. Current tools are limited to very detailed, data-hungry mechanistic models or very general regression models. There are numerous problems with the regression models that cause many to question whether they are appropriate for projecting future ranges.I propose an alternative method that contains some mechanism but is readily applied across many species. Specifically I explore how species abundance varies with a single climatic factor. I then use Liebig's law to combine predictions from different single climatic factors. I test the performance of this model relative to current regression approaches.
Results/Conclusions
Both birds and trees of North America show a pattern where abundance has an upper constraint that is Guassian-bell-curved in shape but with many sites falling below the upper constraint. This confirms as general a pattern previously documented only in a single species (beaver or Castor canadensis). I show that combining this pattern with Liebig's law produces a predictive tool. For predicting modern ranges from modern data, this tool makes moderately accurate predictions (as measured by r2) but which are less accurate than regression methods. I then compare the accuracy of this tool relative to regression methods for making projections into the future using paleontological data.
