Read our COVID-19 research and news.

Downfall. This graph shows the 1992 North Atlantic cod fishery crash.

Millennium Ecosystem Assessment of the United Nations

Predicting Collapse

In 1990, North Atlantic fishers hauled in more than 200,000 metric tons of cod; in 1992 they caught almost none. The collapse cost thousands of Canadian fishers and plant workers their jobs, and the northern cod fishery has never recovered. Now, physicists studying laboratory yeast have found a new way to tell when such a collapse is imminent. The researchers hope their warning signal can help fishery and wildlife managers act in time to save stressed populations.

The team's work is "a really nice paper" that "could potentially lead to some new insights," says Stephen Carpenter, an ecologist at the University of Wisconsin. Madison, who has studied similar early warning signals in lakes.

The key to preventing a population collapse is spotting early signs of trouble. One recognized warning signal is that unhealthy systems often take longer than healthy ones to recover from a disturbance. Scientists call this "critical slowing down." For example, Carpenter and colleagues found that algae levels were slow to return to normal in a lake to which they had added largemouth bass, a predatory fish. However, measuring this slowing in nature can require years' worth of data—a luxury that many fishery and wildlife managers don't have.

Enter Saccharomyces cerevisiae, commonly known as brewer's yeast. In June 2012, physicist Jeff Gore of the Massachusetts Institute of Technology in Cambridge and colleagues reported that they had induced critical slowing down in laboratory populations of these single-celled fungi. Brewer's yeast cells break down inedible sugars in their environment into edible ones, meaning that individuals get a boost from the work of their neighbors—especially at high densities. Thus, the scientists were able to stress their populations by diluting them. The researchers found that at lower densities, the populations took longer to return to previous levels after being shocked with a dose of salt—which can harm or kill yeast—in their growth medium. But an isolated lab colony is a highly artificial system. In real ecosystems, creatures can move from one part of their environment to another.

So, in their new study, Gore and colleagues connected yeast populations by migration. The researchers grew groups of yeast colonies in rows of small, circular wells on plastic trays—imagine the bottom half of a miniature egg carton that has eight rows instead of two. Every morning, the scientists transferred a quarter of each population to the wells on either side of it, simulating a natural dispersal process like fish swimming from one region of a lake to another. The team then diluted the cells in one of the wells in each group to an extremely low density, creating what they call a "bad region." The researchers measured each population's size over the next week and found that colonies to either side of the stressed yeast in the bad region also declined . But populations two or more wells away remained healthy.

Then the scientists delivered the knockout blow. They began diluting yeast cells in the rest of the wells, bringing the entire group closer to collapse. They found that the distance between the bad region and the nearest well with a healthy population increased from two wells to three or more. Gore and his colleagues call this distance the "recovery length" and believe it could be observed in real-world environments with habitats of different levels of fitness. For instance, managers could monitor fish numbers in a marine reserve next to a fishery and curtail fishing if the distance from the fishery's edge to the nearest region of healthy fish populations increases. Recovery length is a "new category of indicator that has not been proposed in the field," says Gore, whose team published its results online today in Nature.

"The really cool thing about the insight is that it could be applied in field conditions," Carpenter says. He envisions ecologists studying satellite images of rangelands or other ecosystems to look for increases in recovery length. Carpenter also notes that while "real landscapes are far more complex than a one-dimensional gradient of yeast cultures," Gore's lab experiment adds something that whole-ecosystem studies like his often can't: easy replicability and control. "The yeast experiments give us one more angle to think about the problem," Carpenter writes in an e-mail.

For Gore, the next step is reaching out to people who could benefit from his results. He has recently started collaborations with scientists who work on economically and ecologically important systems like fisheries and honey bee colonies. "Nobody wants to necessarily save my microbial populations from collapsing," he admits. "We really want to save populations in the wild."