Read our COVID-19 research and news.

When forests of C. amentacea algae are degraded, they soon fall victim to algae that harbor less biodiversity.

Benedetti-Cecchi and Luca Rindi/University of Pisa

How just one data point could predict the collapse of an entire ecosystem

Catastrophic ecosystem failures can destroy industries and threaten food supplies—just think of the collapse of the North Atlantic cod fisheries in the early 1990s or insect outbreaks that devastated Northern spruce and fir forests in the 1970s. Now, a new study promises a way to predict—and possibly head off—the collapse of species that support those ecosystems, with just one measurement.

For years, scientists studying lakes and other ecosystems have predicted collapses before they happen, using long-term data on environmental factors and population health. But scientists want warning signals that don’t require years of expensive monitoring. In a 2013 laboratory study, physicist Jeff Gore of the Massachusetts Institute of Technology in Cambridge took a step in that direction by predicting when yeast colonies were about to die of stress because of low population densities. The longer the distance between healthy and unhealthy colonies, the likelier they were to collapse. Gore and his colleagues named this measure “recovery length.”

To see whether the measurement would hold up in the field, Gore teamed with Lisandro Benedetti-Cecchi, an ecologist at the University of Pisa in Italy, to study algae along the shore of the small Italian island of Capraia. They created 2-meter-long study plots with adjacent areas of two types of algae: ecosystem-supporting miniforests of greenish brown Cystoseira amentacea and turflike species that nurture far less biodiversity. The researchers then removed different fractions of C. amentacea from each plot—0%, 25%, 50%, and 75%. The more they removed, the farther the turf algae invaded. By measuring how far the turfs invaded, the researchers determined each plot’s recovery length. After 2 years, in the plots missing 75% of the forest algae, the ecosystem tipped over to entirely turf algae and C. amentacea never came back, showing that recovery length could predict ecosystem collapse, the researchers reported yesterday in Nature Ecology & Evolution.

Gore and Benedetti-Cecchi say that recovery length could indicate impending shifts in other important ecosystems. Gore dreams of testing it in fisheries where protected areas abut heavily fished ones: If the method works, he hopes fishery managers can use it to set catch limits to avoid a collapse. Stephen Carpenter, an ecologist at the University of Wisconsin in Madison, is also excited about the method’s potential to quickly signal impending collapses. For instance, satellite imagery could be used to test whether it predicts the collapse of rangelands in the western United States and other places because of invasion of weeds or desertification, he says.

But it won't be easy or even possible to measure recovery length in all ecosystems, researchers warn. For instance, Carpenter is unsure whether it would be useful in the lakes he studies, where waters of varying quality tend to mix quickly.

Still, demonstrating the principle outside the lab is a valuable step, says Marten Scheffer, an ecologist at Wageningen University in the Netherlands. “It is good to see the work now taken to the great outdoors.”