Yale Graduate Challenges Existing Ideas About Ecosystem Models

Yale Ecology and Evolutionary Biology Professor Michael Donoghue is a phylogenetic biologist who studies the evolution of plants. Recent research from his lab, published in the Proceedings of the National Academy of Sciences,  poses a challenge to current ideas about ecosystem modeling that assume that individual plant functional traits can be used to make predictions about ecosystem properties such as biomass and productivity.

Dr. Elisabeth Forrestel GRD ‘15, lead author of the study, found that there were fewer grass species in the Midwestern region of North American compared to those in South Africa, and that these regions also possessed a distinct set of grass lineages across otherwise similar participation gradients.

“The point is that on these different continents, grasslands look similar, but upon closer analysis there are striking differences in composition,” Donoghue said. “In particular, at the wet end of the spectrum, they’re both dominated by similar grass types, but at the dry end there are very different groups that dominate.”

In the study, Forrestel shows that the grasses in South Africa and North America have different functional traits (such as the size of the stomata through which gases are exchanged), and that these traits exhibit different patterns across the precipitation gradient.

In spite of these differences, Forrestel found that the two regions were nearly indistinguishable in terms of the relationship observed between precipitation and overall productivity. “This result is a bit surprising because we’ve always thought that these functional traits in particular would predict how the ecosystem would behave,” Donoghue said. “We thought we could make accurate predictions about how a characteristic like plant height would influence the productivity of the system, but it’s more complicated than this.”

Forrestel’s study highlights a missing link in the development of trait-based approaches to modeling ecosystem function. She challenges the notion that there exists a single set of traits that allows plants to function at higher level in certain environments. Her results suggest that scientists can no longer look at a specific trait in isolation when modeling ecosystem function but must also consider the role of different lineages of plants in different areas.

“You really have to take into account a lot more information to make a good model,” Donoghue said. “What we’d like to do is to develop a better global modeling framework where we can make more accurate predictions.”

Source: Forrestel, EJ et. al. “Different clades and traits yield similar grassland functional responses” Pro Nat Acad Sci. Vol 114 No. 4. 705-710