Using fossil plants to more accurately understand past climates

December 28, 2016 | Ethan Hyland

I spend much of my time in the field, exploring new rock outcrops in Argentina, searching for new fossils in Saudi Arabia, and trying to find records of past life we haven’t yet discovered all over the world.

But this year, I realized that there was plenty to discover right here in Seattle at the Burke Museum! The Burke’s paleontology collection contains more than three million fossils containing valuable scientific insights, but many of which have yet to be studied in detail.

So I dove into the Burke paleobotany collections, searching for plant fossils that I could use to solve a question that had been frustrating my research group for a while: why don’t temperature estimates from leaf fossils in the western United States agree with each other?

You may not have noticed this by looking at leaves around Seattle, but in colder places leaves more commonly have teeth along their edges, while in warmer places they tend to have smooth edges. People have measured this phenomenon (and other leaf shape and size characteristics) in floras worldwide and created a group of mathematical relationships that can actually estimate the temperature of a place where any given set of plants grew.

The problem is, if we use these modern relationships to predict temperatures from fossil leaves in the western United States, they all give different answers!

A view of two leaves one pointy with smooth edges and one round with toothed edges
"Photo: Jack Wolfe "
"Photo: Jack Wolfe "

Example leaves from cold and warm floral assemblages, and the linear relationship between mean annual temperature and the proportion of entire margin (smooth) leaves.

A chart showing the variation of temperature

Wolfe, J.A., 1979, Temperature parameters of humid to mesic forests of eastern Asia and relation to forests of other regions in the Northern Hemisphere and Australasia: USGS Professional Paper 1106, p. 1 - 37.

Our plan was to use a brand new method for estimating temperatures called “clumped isotopes,” a method partially pioneered in the Huntington lab at the University of Washington, to independently figure out which of these many plant-based methods would get us the right answer. This new method doesn’t rely on the plant fossils themselves to calculate a temperature, but instead uses the composition of different carbon dioxide molecules within the rocks that preserve those fossils.

Once I had located all the right fossils in the Burke’s paleobotany collections, I sampled the material around them with a tiny drill and then used the equipment in UW’s IsoLab to make new temperature estimates for the fossils.

The results were exciting! Not only could we resolve the issue of why these different plant-based methods didn’t give us the same answer, but we could also use all these robust new temperatures to say something about the paleoclimate of the western U.S.

Our comparison allowed us to choose which plant-based method works best in North America, which lets us estimate past temperatures in places where we have plant fossils but no carbonate-bearing (limestone) rocks. This opens up possibilities for more researchers to answer research questions using the plant fossils in the Burke collection!

But that wasn’t all that we learned. The new temperatures from all of these sites also told us something about past environments in the western U.S., particularly in the places where we had lots of samples like on the Colorado Plateau.

We noticed that the oldest samples (50 million years old) from the northern part of the plateau (Colorado, Utah) showed warmer temperatures than samples from the southern part of the plateau (Arizona, New Mexico), while the youngest samples (25 million years old) showed the opposite pattern.

As Seattleites know well, the higher you go in elevation the colder it gets, so what this new data told us was that the northern part of the plateau started out lower in elevation than the southern part, but over time grew to be higher. This makes sense if we think about modern elevations in the west, but it suggests something new about this area further back in the past—specifically it suggests that the Colorado River used to flow north instead of south.

This is an idea that geologists have argued about for many years, but elevation estimates from the plateau had previously been missing or confusing. Hopefully our new measurements will help put the debate to rest. Look for our results in an upcoming edition of the journal Earth and Planetary Science Letters!


Ethan Hyland is a University of Washington Future of Ice Postdoc, working with the departments of Earth & Space Sciences and Biology to study past climate and ecology in unique environments worldwide.  His co-authors on this study are Kate Huntington, Associate Professor in the Department of Earth and Space Sciences, UW, and Nathan Sheldon, University of Michigan.