To determine the effects of climate change on biodiversity, researchers often rely on modelling distributions or niches of species using a combination of occurrence records and climatic data. Such models might predict that the range of a species is likely to contract under future climates, but they may not be able to capture local adaptation.
In this study, researchers used genetic information to define ecological subregions, downloading GBIF-mediated occurrences of Fremont cottonwood (Populus freemontii) and modelled them separately by subregion. The models showed that genetically distinct ecotypes differed significantly in climate niche space, showing adaptation to local environment. The models were better able to predict populations than those that did not consider the genetic component. When addressing impacts of climate change, the researchers found that ecotypes varied greatly in their response, as some were predicted to gain more than 60 per cent in suitable habitat while others stood to lose it all.