Early Warning for Endangered Species?

Ecologists have long wondered how to tell when a population of animals is about to slide toward extinction. A new study provides a statistical alarm that does just that. "The paper is an important test of early-warning signals," says ecologist Stephen Carpenter of the University of Wisconsin, Madison, who was not involved in the work. The result was based on a simple laboratory experiment, but if the technique works in the wild, it could help land managers know when they need to step up their conservation efforts.

The major causes of animal extinctions today are clear: habitat loss or degradation, invasive species, hunting, and pollution. But it's been nearly impossible to predict whether or when an endangered population will become so imperiled that it is doomed to vanish if nothing is done to help. Theoretical biologists have suggested that this tipping point toward extinction would be preceded by a phase of so-called critical slowing down, a mathematical phenomenon seen before a variety of events, including abrupt change in past climates. For endangered species, critical slowing down means that a population takes longer and longer to bounce back from small declines.

John Drake, an ecologist at the University of Georgia, Athens, and Blaine Griffen of the University of South Carolina, Columbia, decided to look for signs of critical slowing down in the lab. Griffen set up 60 small tanks and put populations of water fleas (Daphnia magna) in each. After letting the Daphnia get settled for several months, Griffen mimicked environmental degradation in half the tanks by gradually cutting back on the blue-green algae he fed them. These nutrition-deprived populations began to slowly shrink. After about 270 days, they passed a tipping point and died out within a bit longer than a year. The other 30 populations, which received a stable food supply, fluctuated in size but persisted.

Drake then analyzed the trends in population size. Four statistical measures of abundance showed signs of critical slowing down in the degraded environments but not in the healthy ones, Drake and Griffen report online today in Nature. The slowing down lasted as long as eight generations before the populations crossed the tipping point to extinction. The experiments offer "remarkable statistical resolution" of critical slowing down, says William Fagan, a biologist at the University of Maryland, College Park.

"It is indeed a milestone article," adds Marten Scheffer of Wageningen University in the Netherlands. "It provides experimental evidence for generic early-warning signals that have been theoretically predicted." One advantage of looking for such warning signals, Drake adds, is that it would not require detailed knowledge of the species, such as the amount of habitat needed, only a comparison group or baseline data on population trends. In principle, that could make critical slowing down a broadly applicable tool for conservation biology.

The big question is whether the signal can be detected in noisy ecological data from the real world. For the species of highest concern, Fagan notes, the population estimates are much scarcer and more uncertain than the laboratory data that Drake and Griffen analyzed.