Predicting species interaction and population dynamics in ecological communities

By Peter Murphy

Published January 16, 2020

“Different species of biological organisms like plants, animals and bacteria interact with each other in mysterious ways,” says Zhenduo Zhu, an assistant professor in the Department of Civil Structural and Environmental Engineering at UB, “Would it be possible to forecast changes in population dynamics in the near future, and how does an invasive species perturb the interaction of resident species?”

Relationships between species

“The new method will address the challenges of noisy macroscopic ecological data.”
Zhenduo Zhu, Assistant Professor
Department of Civil, Structural and Environmental Engineering

Zhu, Chen Liao, a postdoctoral scholar and Joao B. Xavier, a faculty member of the Computational and Systems Biology Program, Memorial Sloan Kettering Cancer Center, answer these questions with their latest published article, Enhanced inference of ecological networks by parameterizing ensembles of population dynamics models constrained with prior knowledge.

Zhu and his colleagues studied the ecological consequences of Asian Carp invading the Illinois River, but the team needed more data than originally anticipated, according to Zhu. “We soon realized that these questions cannot be answered without knowing how different species within the same environment interact with each other,” Zhu says.

The researchers applied the generalized Lotka-Volterra (gLV) equation to a study on the freshwater fish ecosystem in the Illinois River. Liao and Xavier have experience using gLV to model population growth of species in the microbial community.

The researchers applied the generalized Lotka-Volterra (gLV) equation to infer interactions among fish species in the Illinois River. “Compared to statistical models that explain associations between species abundances, gLV is mechanistic and allows us to examine how population growth of a fish species is influenced by others,” Zhu says.

“This new method will address the challenges of noisy macroscopic ecological data,” Zhu says, “gLV allows us to examine population growth of any species as a linear combination of influences from other species.”

The team leveraged fish abundance data from the Long-Term Resource Monitoring Program in the Upper Mississippi River System, one of the few monitoring programs in large rivers in the U.S. to make train the gLV model.

“We developed a new method to parameterize gLV equation from macroscopic ecological field data,” Zhu says, “the challenge for us and the whole field is, ecological data are noisy, sparsely sampled and lacking replicates.”

Using the trained model, Zhu, Liao and Xavier found a nontrivial positive effect of emerald shiner on channel catfish. They predicted that the dominant fish species in the river will continue to fluctuate over the next 20 years. The model also helped the researchers reveal the impacts of invasive silver carp on native species in the ecosystem.

“The invader species has a stronger impact on native predators than on prey,” Zhu says, “this suggests that the invader may perturb the native food chain by replacing the diets of predators.”