'Epidemiological' Study Sheds New Light on Climate Change's Impacts on Forests

April 3, 2011

Tim Lucas, 919-613-8084, tdlucas@duke.edu

DURHAM, N.C. – A painstaking, 18-year study of 27,000 individual trees by Duke University researchers finds that tree growth and fecundity – the ability to produce viable seeds – are more sensitive to climate change than previously thought.

The study, published April 5 in Global Climate Biology, identifies earlier spring warming as one of several overlooked factors that affect tree reproduction and growth, and can help scientists and policymakers better predict which species are vulnerable to climate change, and why.

It also identifies summer drought as an important but overlooked risk factor for tree survival and fecundity, and finds that species within four broad genera of trees – pinus (pine); ulmus (elm); fagus (beech) and magnolia – are particularly vulnerable to climate variables.

“In a sense, what we’ve done is an epidemiological study on trees to better understand how and why certain species, or demographics, are sensitive to variation and in what ways,” lead author James S. Clark says of the project.

Clark is H.L. Blomquist Professor of Environment and professor of biology and statistics at Duke’s Nicholas School of the Environment.

To conduct their study, he and his colleagues measured and recorded the growth, mortality and fecundity of each the 27,000 trees at least once every three years, ultimately compiling an archive of more than 280,000 tree-years of observed data.  Using a specially designed bioinformatic analysis, they quantified the effects of climate change on the species over time, with spatial climate correlations. This approach allowed them to calculate the relative importance of the various factors, alone and in combination, including the effects of localized variables such as competition with other trees for light, or the impact of summer drought.

“As climate continues to change, we know forests will respond.  The problem is, the models scientists have used to predict forest responses focus almost solely on spatial variation in tree species abundance – their distribution and density over geographic range,” Clark explains.

If all trees of a species grew in the same conditions – the same light, moisture, soil and competition for resources – this generalized, species-wide spatial analysis might suffice, because scientists wouldn’t need to worry about demographic variables and risk factors when trying to predict biodiversity losses due to climate change. “But in the real world, we do,” Clark says. “That’s where this new concept of climate and resource tracking of demographic rates comes in.

“Trees are much more sensitive to climate variation than can be interpreted from regional climate averages. By quantifying the effects and relative importance of competition and climate variables, including previously overlooked impacts on fecundity, over both time and space, the model we’ve developed addresses this need,” he says, “and can be used to help guide planning.”

The trees studied included 40 different species, located in eleven different forest stands in three geographic regions of the Southeast – the southern Appalachians, the Piedmont and coastal plain.  They were subjected to both natural and experimental variations.

Grants from the National Science Foundation funded the study.

Clark’s co-authors are doctoral student David M. Bell and research associate Lauren Nicholas, both of Duke’s Nicholas School, and Michelle H. Hersh, who received a PhD in ecology from Duke in 2009 and is now a postdoctoral researcher at Bard College and the Cary Institute of Ecosystems Studies.