Fluctuations in Fish Diversity Affect Ecosystem Stability

A new study of a community of fish species confirms the theory that more species-rich ecosystems are also more stable. More surprisingly, it also suggests that both diversity and stability change far more frequently than previous studies had recognized. This finding could lead to methods that better anticipate, and help managers prevent, sudden ecosystem collapse.

The idea that ecosystems with a greater diversity of species are more stable is an old one in ecology. Numerous studies have supported it, but typically by comparing the stability of systems with low diversity to those with high diversity, based on the assumption that diversity and stability are relatively constant through time.

“This paper is different because it not only shows that diversity begets stability, it looks at how stability itself changes over time within a single ecosystem,” said George Sugihara, a study author and biological oceanographer at the Scripps Institution of Oceanography. “The fact that there are episodes of high and low diversity and stability is a game changer for how we view ecosystems.”

The research appears in the journal Nature and was supported in part by the Lenfest Ocean Program.

The study was possible because of a rich dataset from Maizuru Bay, near Kyoto, Japan. Reiji Masuda, another study author and professor of fisheries at Kyoto University, undertook scuba dives every two weeks for 12 years. He recorded the size and identity of all the fish he encountered. This frequency and duration of sampling is unusually high for field studies.

From these data, the authors were able to determine how species diversity changes over short time scales. However, it was more difficult to quantify stability and how it is affected by diversity. Ecologists have developed various ways to measure stability, but they apply best to ecosystems that are in equilibrium. But recent studies suggest that most ecosystems do not have a stable equilibrium state, which would make it difficult to know whether variability is a sign of instability or just normal ecosystem dynamics.

In this study, the authors argued for defining stability as the ecosystem’s ability to bounce back from perturbation. To measure this “dynamic stability,” they turned to a powerful mathematical approach known as empirical dynamic modeling, or EDM, being developed at Scripps.

As a first step toward calculating stability, the researchers looked at 14 key fish species. They used an EDM technique called convergent cross mapping to determine whether species were affecting each other’s abundance. (For example, increasing the abundance of one competitor can decrease the abundance of another competitor.)

Next, they looked at how these interactions between species changed over time. In fact, they looked at the combination of all 14 of the significant interactions simultaneously using a mathematical matrix—a 14-by-14 grid of numbers defining the strength of each interaction. (This matrix was described in a publication 2016 paper by Ethan Deyle, a postdoctoral researcher in the Sugihara laboratory.)

From this matrix, which also changes over time, the researchers calculated a mathematical indicator of system stability (known as a “local Lyapunov exponent”). This measures how sensitive the system is to small perturbations.

The final step was to assess whether stability was influenced by diversity. The researchers found that it was; specifically, high diversity in summer led to high stability, whereas low diversity in winter led to low stability. But the diversity-stability link was indirect. The more direct cause of high stability was the fact that when diversity was higher, species interactions were weaker on average, so that a perturbation of one species was less likely to greatly affect others.

The episodic shift in diversity and stability suggests that the common practice of monitoring ecosystems annually may not be sufficient. For example, annual monitoring might fail to detect episodes of low diversity and stability that could trigger an unwanted outcome.

“This method has the potential to provide warnings of sudden changes in many ecosystems,” said Hao Ye, a study co-author. “This would allow resource managers to better anticipate and respond to events like population collapse or invasive species.”

The study is in part the result of a project, supported by the Lenfest Ocean Program, to tap into the power of EDM to aid in fisheries management. Additional results are expected in February or March 2018, on the subject of Atlantic and Gulf menhaden.