Award: DBI-1759759

Earth's climate impacts the distribution of species, their abundance, and how they form communities. With past climate changes, some species have shifted their distributions accordingly; other species went extinct. To understand how species can and will respond to environmental change, we can study how they have responded in the past. A period of rapid climate change, before anthropogenic warming,  occurred after the last ice age about 21,000 years ago to the current “interglacial” state. During this time, many species in North America moved northward, tracking glacial retreat. Scientists have a general understanding of how species moved as these changes occurred, but many open questions remain. These include: Where did species live during the Ice Age? When did they start to move from these places?  What pathways did they take when they moved across the land, and how fast did they move?  Did species that occur together now migrate together at the same time and pace? Scientists have used fossils, genetic data, and modern species distributions to address these questions. However, the inference made about species movements differs among data types. Each data type tells us something about past species movement, but each has limitations. Our work used these multiple lines of evidence in a coherent analytical framework to address these longstanding questions.

We developed a modeling framework that combines: pollen records, which are deposited in the beds of lakes and so provide evidence of the presence of particular trees in specific places; genetic data, which can indicate how many places species survived in during the glacial period; and spatial location of modern herbarium specimens which can be  used to infer in what climates the species would have preferred in the past. This framework involves simulations that describe different scenarios of how species moved across the land, and how the influence of these movements on patterns of genetic variation over time. We also used pollen data and modern herbarium data, coupled with climate data, to constrain the simulated scenarios. We then compare observed and simulated datasets to identify the set of scenarios that best recreate patterns we see in modern genetic samples. These select simulations form our best estimate about how the species moved across the landscape over the last 21,000 years. 

For our work, we focused on green ash, a species that occurs across eastern North America but is currently imperiled by the emerald ash borer, an invasive species. Our results uncover important differences among our estimates of how fast green ash moved as glaciers retreated. Combining different data types notably reduced uncertainty in our postglacial movement estimates. This underscoring the importance of combining data. Our results also revealed critical information about the locations where green ash persisted during the last ice age, with important implications for the conservation of the species. Specifically, we found that green ash likely survived in three separate locations, two of them in the south and one close to the edge of the ice. We also found that green ash probably moved northward up to several hundred feet per year by dispersing their seeds to uninhabited locations. Our software and methods are described in publications and freely available for other scientists and interested people to peruse and use.

We developed a set of outreach materials based on our research. First, we created a set of lesson plans for primary, secondary, and college-aged students for understanding how scientists predict where species were in the past and could be in the future under changing climates. Two groups of teachers and one class of college students engaged in these lessons. We also translated from English to Spanish in-depth instructions from a scientific workshop on how to model species as climate changes. Across the years of our work, we trained 2 early-career scientists and 5 female undergraduate, recently graduated students, or master’s students. Finally, we held two workshops for almost 50 scientists to better develop ways for combining different types of data for understanding how environmental change affects the biosphere.

Last Modified: 10/15/2023
Modified by: Sean Hoban