2021 Hertz Fellow Emily Geyman studies perturbations to Earth’s climate across history in order to better predict the climate response to greenhouse emissions.
Geyman has spent the last two years living in the Arctic Circle, Norway, as a Daniel M. Sachs Global Scholar at Princeton University. She received a bachelor’s degree in geosciences from Princeton University in 2019 and will begin her PhD at California Institute of Technology in fall 2021.
On her current research:
I’m living and working in Svalbard, which is an archipelago in the Artic Circle, home to a little more than 1,500 glaciers. These small glaciers and ice caps are contributing to sea level rise just as much as Greenland and significantly more than Antarctica. So understanding what the controls are on how quickly they’re melting is really important.
In the 1930s, there was a Norwegian expedition to map Svalbard and they took aerial images that cover almost the whole archipelago. Those were supposed to be used to make high quality topographic maps, but World War II postponed that project and those images basically got left in the closet for 80 years.
I’ve been using structure from motion photogrammetry to turn those old images into 3D models of Svalbard’s ice surface in the 1930s. From that we can compare it to today and look at which glaciers are shrinking and by how much, and start to understand how quickly they’re melting and make better predictions for the future.
On the allure of geoscience:
One thing that’s special about geoscience is that the laboratory is basically the world that you live in. Once you start taking geology and geomorphology classes, looking out of an airplane window is never the same again.
On the future of her field:
Machine learning, deep learning, and computer vision are permeating every field right now. The thing that makes the application to Earth science really different is that there’s a hurdle to being able to take advantage of all of these new computational resources and techniques, which is figuring out how to make the inputs quantitative.
Geological observations are quite rich, but they’re not really ready to be analyzed in the ways a lab experiment would be. I think that first step is a really exciting extra challenge. Once we find more intuitive, efficient, and effective ways of translating these geological data sets into data streams, then we can start taking advantage of these better tools for figuring out what has happened, what are the climate signals in Earth’s past, and what they can tell us about future climate.