The ecology of geographic range limits.

Identifying processes underlying species' geographic range limits is currently at the forefront of ecological research, in part due to increasing attention to climate change effects on the distributions of organisms. Our understanding of range limits has benefitted from a rich body of theory, but several influential reviews have insisted that we know little about causes of range limits in nature because there is scant empirical work relative to theoretical. More recent syntheses show mounting support for different hypotheses, although they are not always able to separate different ecological processes. In this review, I recommend a shift in how researchers think about range limits: (i) identifying a first-principles hypothesis of range limits which should structure range limit studies; and (ii) reconsidering the processes that limit geographic distributions, which are relatively few and redundant. First, I argue that estimating the scenopoetic niche (habitat requirements that exclude biotic interactions) allows a first-principles approach to understanding geographic distributions and limits. Some general empirical support for the scenopoetic niche as a primary range-limiting factor has accumulated. Estimates of the scenopoetic niche will structure subsequent tests of range-limiting processes based on how it underpredicts or overpredicts species' distributions. I discuss observational and empirical ways of testing whether the scenopoetic niche is actively limiting species' distributions. Second, I review various theoretical models of geographic range limits; theoretical ecological models only vary a few key parameters, so our understanding of what limits species' geographic distributions might be much better than previously asserted. For instance, predation, competition, and parasitism are all distinct biotic relationships, but they are all antagonistic biotic interactions that can influence range limits in similar ways by reducing the set of conditions under which a species could persist. Several issues complicate causal inferences from static geographic patterns that have remained problematic in empirical work for decades. These issues are related to spatial autocorrelation and interpretation of range overlap. Theoretical metapopulation models have been developed to understand how range limits can form, although metapopulation processes are understood to be consequences of ecological dynamics that are formally modelled in non-metapopulation models (e.g. per capita effects of interspecific interactions, effects of the abiotic environment). I then discuss methods for empirical tests of various range limit hypotheses. Implementation of different methods will depend on tractability with geography and ecology - many researchers cannot survey remote areas, study demography of long-lived organisms, collect large sample sizes for rare species, or conduct field manipulations. However, at least some of the methods showcased will have applicability for any study system. Finally, I suggest a research agenda for improving our understanding of the ecology of geographic range limits: better teasing apart of ecological causes of range limits (e.g. scenopoetic niche versus various biotic interactions) and incorporating spatiotemporal variability. Hopefully some of the ideas highlighted here will support more conservative interpretations of data, stimulate tests of multiple hypotheses, and produce better predictions of changes in geographic distributions under environmental change.