Studies seeking to predict geographic range changes under climate change frequently use ecological niche modelling (ENM), relating current species distributions to climate variables and projecting the models into future climate scenarios.
Recognizing that the recent past does not offer reasonable analogs for the unprecedented CO₂ levels predicted in the next century, this study of turtles used a deep-time record and explored 100 million years of paleoniche dynamics.
Using species occurrences of extant terrestrial and freshwater turtles from GBIF and fossil records of extinct relatives from the Paleobiology Database, the authors created ENMs, leveraging three modelling algorithms combined in an ensemble approach.
They projected these models into three past hyperthermal scenarios (38–34, 72–66 and 94–84 million years ago) as well as future predictions based on several Representative Concentration Pathways (RCP) emission scenarios.
Relying in insights from the range shifts observed in the deep past scenarios, the authors showed future predictions of poleward range expansions in the Northern Hemisphere with decreased habitat suitability at lower latitudes. This pattern represent an inversion of trend otherwise present since the Eocene—34 million years ago.
The authors note that while some freshwater turtles may able to track their changing niches better than terrestrial species, habitat destruction and fragmentation may limit the ability of both groups to cope with future climate change.