When George Sugihara reads about credit crises and federal bailouts, he is inclined to think about sardines—California sardines, to be precise.

A few decades after the Great Depression, the sardine fishery in California was suffering from a similarly devastating collapse. Fishers who had generally landed more than 500,000 tons of sardines annually during the 1930s caught fewer than 5,000 tons during the worst years of the 1950s and 1960s. Whereas a few Cassandras might have warned of trouble in each case, nobody could have predicted exactly when each collapse would come or how severe it would be.

The sardine collapse puzzled fisheries experts. Some blamed overfishing. Others suspected environmental swings—shifting wind patterns or cooling sea-surface temperatures. But nobody could prove either case. Eager to prevent another such collapse, California set up a monitoring system that has been collecting data on sardine larvae for the past 50 years. Sugihara, a mathematician and theoretical ecologist at the Scripps Institution of Oceanography in La Jolla, Calif., analyzed that data and came to a surprising conclusion: both potential explanations of the sardine collapse were wrong.

His conclusion, in a study published in Nature in 2006, was that the problem was the harvesting of too many big fish. Fishing boats were leaving behind a population of almost all juveniles. Sugihara showed that mathematically such populations are unstable. A slight nudge can create a boom—or a catastrophic collapse.

Imagine, Sugihara says, a 500-pound fish in an aquarium. Feed it more, and it gets fatter. Feed it less, and it gets thinner. The population (of one) is stable. But put 1,000 half-pound fish in that aquarium, and food shortages could result in the deaths of hundreds, because the small fry have less stored body fat—and therefore cannot ride out a short famine. Food abundance does not necessarily mean all the fish get bigger, either; it could en­courage reproduction and a population boom—which might in turn overwhelm the food supply and lead to another bust. It is an unstable system. “That’s the reality of fisheries, of economies, of a lot of natural systems,” Sugihara says. The recent history of the sardine fishery illustrates that instability: fishers along the West Coast from Canada to Mexico are now harvesting a million tons of sardines annually.

But this instability is not understood by people who run fisheries, Sugihara insists. By law they manage fisheries for “maximum yield.” The notion that such a maximum yield exists implies that fish grow at an equilibrium rate and that the harvest can be adjusted in accordance with that growth to keep yields stable. In contrast, Sugihara sees fisheries as a complex, chaotic system, akin to financial networks. They are so alike that the global financial giant Deutsche Bank lured Sugihara away from academia for a time; there, from 1996 to 2001, he successfully used the analytical techniques that he would later call on for his sardine work to make short-term predictions about market fluctuations.

Although both marine ecosystems and financial markets might look random, Sugihara explains, they are not. That means making short-term predictions is possible, as it is with the weather. The eminent ecologist Robert M. May of the University of Oxford calls that predictability “the flip side of chaos.” May oversaw Sugihara’s doctoral work at Princeton University when he was a visiting professor there and is now a frequent collaborator. “George was one of the first to see this as a recipe for making predictions,” he says.

Sugihara’s research comes at a time of enormous concern about the future of the world’s fisheries. Perhaps the most alarming report came in late 2006, when Boris Worm, a marine conservation ecologist at Dalhousie University in Nova Scotia, reported in Science that for 29 percent of currently fished species, the catch had dropped to less than 10 percent of the historical maximum. If the trends continue, he reported, all fisheries around the globe will collapse by 2048.

Others think the future is not nearly so gloomy. “It’s very dependent on where you are,” comments Ray Hilborn, a professor of fisheries management at the University of Washington. The U.S., Canada and some other developed countries have cut fishing rates, and the future looks brighter, he says. But Asia and Africa lack effective fisheries management, and even European countries have failed to agree on solid management plans. Fisheries in those regions are in far greater peril, Hil­born states.

The practical implications of Sugihara’s work are clear. Current fishing regulations usually have minimum size limits to protect smaller fish. That, Sugihara maintains, is exactly wrong. “It’s not the young ones that should be thrown back but the larger, older fish that should be spared,” he explains. They stabilize the population and provide “more and better quality offspring.” Laboratory experiments with captive fish back up Sugihara’s conclusions. For instance, David Conover of Stony Brook University found that harvesting larger Atlantic silversides from his tanks over five generations produced a population of smaller individuals.

Sugihara has also shown that populations of different fish species are linked. Most regulations consider each species—sardines, salmon or swordfish—in isolation. But fishing, he says, is like the stock market—the crash of one or two species, or a hedge fund or mortgage bank, can trigger a catastrophic collapse of the entire system.

Sugihara has also used his combined experience in ecology and finance to work on new kinds of fisheries management schemes. One is the notion of tradable “bycatch” credits. Bycatch refers to the turtles, sharks and other animals that fishing fleets do not seek but catch accidentally. In the tradable bycatch credits plan, fishing boats could be allocated a certain number of credits. As they used those credits, they would need to stop fishing or buy more credits on the open market. As the bycatch increased, the number of outstanding credits would fall, and their price would increase. Fishing boats would thus have financial incentive to minimize their bycatch—because by doing so, they could keep fishing longer.

Sugihara’s work on fisheries has not met with universal acceptance. Roger Hewitt, assistant director of the National Oceanic and Atmospheric Administration’s Southwest Fisheries Science Center in La Jolla, remarks that Sugihara’s work is “a bit disconcerting” to people in fisheries management. “In fisheries, the classical approach is to model populations based on first principles. We know how fast [individual fish] are growing, how fast they are reproducing, how old they are when they mature, how many babies they have,” Hewitt explains. “George’s approach is an entirely different one. He looks at past behavior to see if he can predict future behavior.” In a crude sense, Sugihara does not need to know about growth rates, reproduction
or mortality.

Barry Gold, leader of the marine conservation effort at the Gordon and Betty Moore Foundation in San Francisco, describes Sugihara’s analytical tools as “important for understanding how we manage fisheries.” But he thinks that Sugihara’s analysis needs a real-world test: “Until it’s in the field and we see how the fishing industry responds to it, we won’t know how it’s going to work.”

Partly in response to Gold and others about the lack of convincing field tests, Sugihara is now negotiating with fishing industry groups to try to put his work into practice. “Once you’ve stopped imagining that the world is a watch, that it’s extremely predictable, you can make relatively good short-term forecasts,” he states. “I have a lot of faith in human ingenuity. But the first step is acknowledging the reality.”

Part of a 12-step program for fisheries ecologists? “Yes,” Sugihara says. “It feels a little bit like that.”

Note: This article was originally printed with the title, "Chaos and the Catch of the Day".