Traditionally, species distribution models (SDMs) are applied assuming occurrence is a binary state: either an organism is there, or it is not. What happens when it’s not that simple?
Sometimes we have evidence of past occurrences, such as old scat, fossilized remains, tree stumps, abandoned burrows, and eDNA, and sometimes we have sites with no evidence of current or past occupation. Sites with pas evidence provide valuable information because they demonstrate that the location at least once was suitable for the species. But using this kind of evidence is hard because we would normally need to match it to the environmental conditions when it was created.
Yes, we can use radio-carbon dating and similar techniques to estimate the age of scat, tree stumps, or other remains, but estimates of age are often wide, making matching them to a specific time slice (and thus to a particular set of environmental layers) challenging. And even if we could get precise age estimates, in many cases we don’t have environmental data that exactly aligns with the estimated age. How can we use evidence of past occurrence to better inform our estimates of species’ niches and past ranges?
In our new paper led by colleague/friend Erik Beever, we show how to use and interpret ordinal SDMs where sites with evidence of past occurrences are matched to present-day environments. Focusing on the American pika (Ochotona princeps), a high-elevation, climate-sensitive species, we show that ordinal models identify winter temperatures as the driving force behind historical extirpations, while binary models (for which we “squashed” no-evidence sites and old-evidence sites into “absences”) identified summer temperatures as more important. A quick summary is provided by Buckrail!
Update: We also made the cover!!!
Combining past and contemporary species occurrences with ordinal species distribution modeling to investigate responses to climate change [open access | journal cover]
Beever, E.A., Westover, M., Smith, A.B., Gerraty, F., Billman, P., and Smith, F. In press. Ecography.
