COD and MAN at Dartmouth

For the past decade} mathematician Dan Lynch has trained his computer models on a swirling enigma in the North Atlantic. At stake: the salvation of one of the world's greatfisheries.

Beginning 25 MILES OFF CAPE COD and stretching northeast almost to Nova Scotia, Georges Bank forms an elevated plateau some 60 to 300 feet below the surface of the Northwest Atlantic. This area, twice the size ofMassachusetts, has for millennia been one of the world's richest ecosystems, its plankton-thick waters supporting tens of millions of cod, haddock, flounder, and other fish. This richness greeted the first European settlers with "an abundance," wrote one Salem minister in 1629, "almost beyond believing." The fish of Georges Bank saved the colonists from starvation, then enriched them, and have played a vital role in New England's culture and economy ever since.

Such extraordinary productivity suggests a benign stability: Yet the Bank's bounty actually arises from a wild and unpredictable complexity of currents. Warm water from the Gulf Stream wells against the Bank's southern flanks. Icy water slices down the Scotian coast from the northeast. Meanwhile, from the Gulf ofMaine, a spinning gyre bumps against and sometimes washes over the Bank's giant, thumbshaped mound. These currents bring the Bank the tons of plankton that make it so rich. Yet they also imperil every fish born here by threatening to sweep it off the Bank's relatively warm, food-rich shelter into the dark, inhospitable open ocean. Such calamities happen regularly.

Scientists recognizing this paradox nearly half a century ago dubbed it the "Enigma of Georges Bank." How did a relatively bare mound scoured by currents that can easily claim ships, much less fish the size of fingernail clippings, become one of the world's greatest fish nurseries?

This question has lately acquired a certain urgency, as three decades of overfishing, first by foreign factory trawlers in the 1960s and then by American boats in the 1970s and 1980s, decimated the fishing grounds. Like several of the world's other major fisheries, Georges Bank is now "severely depleted," its stocks of cod, haddock, and flounder so exhausted that in 1994 the U.S. government closed major sections of the Bank to fishing. With the cod gone and the fishing boats idle, the need to understand the Bank's fundamental paradox became critical.

What logic rules the Bank's chaos of currents? For decades, scientists could only guess. Today, however, Dartmouth professor Dan Lynch is helping explain the enigma. A mathematician working 100 miles inland at Dartmouth's Thayer School, Lynch seems an unlikely solver of ocean mysteries. Yet over the last decade he has begun to define precisely how the currents of Georges Bank make or break fish populations. In doing so he has pushed the limits of computer modeling to help us understand marine ecosystems, helped create a new style of collaborative scientific inquiry, and integrated physics and biology in unprecedented ways to solve environmental problems. And his work offers fish managers and policy makers a crucial tool in their effort to nurse Georges Bank back toward health.

Lynch, himself, doesn't get out on the Atlantic much. Instead, he explores it from his small, rectangular office at 229 Cummings Hall. He keeps the room nautically neat: desk clean, books squarely shelved, credenzas topped with crisply stacked papers. On a table behind his main desk sits the net he casts into the Bank's depths: a rather modest-looking computer terminal cabled to a powerful mainframe computer elsewhere at the Thayer School. With this terminal Lynch manipulates powerful computer models that he's adapted to analyze ocean currents.

Forming the heart of his computer models is the finite element method, or FEM, a mathematical tool developed in the 1960s to study solid mechanics. FEM uses multiple strings of differential equations to model physical processes. To get a mathematical handle on the subject, the method divides the object or area—a bridge, an airplane wing—into a mesh of triangles or (for three-dimensional problems) tetrahedrons of varying size. The mesh allows the model to draw detailed data from the most critical areas (tiny triangles) of study without overburdening it with too much data from the less critical areas (larger triangles). A student might use FEM to describe, say, the vibration of a violin string or the movement of heat through a saucepan; an engineer might use it to predict the fate of a tall building pushed by a strong wind.

When Lynch started using FEM as a grad student at Princeton in the 19705, few had used it to sucessfully study open bodies of water. Despite the method's power, the general feeling was that it couldn't handle the almost infinite variables of openwater systems. But Lynch was young and confident. He was also, he says, a product of his times. Inspired by the country's growing environmental consciousness, he applied the highly abstract tools of his discipline to the wild world of nature's fluid mechanics.

Lynch spent much of the next 15 years applying FEM to ocean currents. After modeling some estuaries and then the English Channel, he had the model ready for the Gulf of Maine, which even then was showing signs of the decline that would culminate in its drastic collapse in the early 19905. The Gulf offered a singularly promising place to study currents. Decades of pioneering oceanographic work had made it the world's most studied piece of ocean and built a broad network of cooperating scientists. This network, formalized in 1991 as GLOBEC (Global Ocean Ecosystem Dynamics) and RARGOM (Regional Association for Research on the Gulf of Maine), would allow Lynch to draw on vast stores of data and expertise and tie his physical studies to biological studies. (Lynch would take a lead role at GLOBEC, serving on its executive committee, and serve as the first director of RARGOM, still headquarted at Dartmouth.)

The Gulf offered currents on which Lynch could float his equations. The upwellings and vortexes set up complex physical dynamics that scientists knew to be biologically critical, but which they still understood only "qualitatively," as Lynch puts it. They could only guess the answer to the Bank's central enigma.

Lynch dove in to find that answer quantitatively. Joining forces with biologists, oceanographers, and scientists at Woods Hole and other institutions, he harnessed information about currents. Lynch's colleagues dropped trackable "drift particles" in the deep basins of the Gulf, where plankton blooms originated, so they could see how plankton moved onto and around the Bank. They studied the way heat and salt are carried both horizontally and vertically. They studied how larvae and very young fish move in different currents. How wind and storms affect the movement of water. It was and is an elaborate, wide-ranging inquiry, involving scores of researchers measuring multiple variables and producing zigabytes of data.

The mesh of Lynch's FEM models essentially netted all this data into a coherent picture of how currents affect fish. Each model used a series of complex differential equations long strings of equations, sometimes taking up several pages that took the data and turned them into a quantified expression of how these processes interacted. One set of equations might describe how Scotian shelf currents affect the gyres within the Gulf; another how those gyres create the conveyor belt that carries plankton to the young cod born on the Bank; another how the timing of the cod's spawning in relation to spring storms affects whether the young cod will remain over the Bank long enough to encounter the plankton.

As they developed the models, Lynch and his colleagues constantly checked the simulations against reality by "backcasting," the reverse of forecasting. They took the raw data from, say, five years previous, fed them into the model, then compared the model's predicted reality against what actually happened in the studied year. Reality thus served as a check on the model and led to improvements in the model's accuracy. In addition, one model would feed another, creating the same sort of interplay that was occurring in the waters over Georges.

To refine the models even faster, Lynch and his fellow scientists around the Gulf of Maine developed a collaborative method he calls "open modeling," in which numerous investigators openly share their data with each other, publishing both the data and the models on World Wide Web pages. Opening the models to a wide range of colleagues allows other researchers to use the data, contribute more data from their own research, suggest improvements, and offer possibilities for connections that the primary researchers might miss. The result, says Lynch, "is a new, group-oriented way of doing science—collaborative computational modeling. It's unlike the old way where a single scientist or small group of colleagues would do an experiment or develop a theory, then unleash it on the world. Lots of your best ideas drain out into this community. But you get a synthesis you wouldn't get for years in the research-and-publish mode."

Over the early 1990s Lynch helped create a digitized version of Georges Bank that quantified some of the ecosystem's most vital dynamics. In 1993 Lynch and seven colleagues described how a dome of relatively stable water close over the Bank protectsmost of the time—eggs, larvae, and young fish. Scientists had previously known something like this was occurring. Lynch's group located, for the first time, the borders between the deadly temperaturedriven advecting currents and the dome of safe water and identified precisely at what depths survival rates increased and by how much.

Then, in a 1994 analysis of winds, currents, and cod populations from a boom year in cod, 1985, and a bust year, 1982, Lynch and his colleagues examined why and how the dome of safe water sometimes broke down. Using current, wind, spawning, and other data from those two years, the team showed how wind-driven ven storms in 1982 dug into the protective dome of calmer water to decimate the young fish. Even though almost five times as many eggs were laid over the Bank in 1982 than in 1985, only a fifth as many cod survived their first year. Had survival rates in the two years been identical, the 1982 class would have produced more than 200 million fish instead of just under ten million. The currents in the windy year had cut the survival rate by roughly 95 percent.

Finally, scientists were offering a quantitative explanation of the Georges Bank enigma. They described how the Bank's currents usually brought more benefit than harm to the fish born there, and quantitatively defined the conditions that helped the fish and those that hurt them. In subsequent work Lynch and his colleagues have refined their models to explore, among other things, just how currents move food to the Bank.

These investigations of the relationship between currents and the biology of fish and their food, says Mike Fogarty, a University of Maryland fisheries scientist who chairs GLOBEC's scientific advisory board, provide just the sort of integrated picture that managers need to revive Georges Bank.

"Dan's work cuts right to the heart of some of our most critical management issues," says Fogarty. His hope, shared by the National Marine Fisheries Service scientists who are trying to balance fishing pressure against the health of the fish, is that these models will soon allow the Fisheries Service to better predict the fate of the fish born in a given year, and thus better manage the fishing pressure on them. Fogarty, like Lynch, says the models need farther refinement before they can be used this way. And it's too soon to know whether even those tools will help restore the incredible vitality that once marked these waters, for that answer will depend on the rest of us as well.

Yet it's clear that Lynch's work has already answered crucial questions about this extraordinaryecosystem's murky dynamics, and that his integrated, collaborative approach holds promise for providing even greater insight. Such insight is badly needed not only here, but elsewhere around the world, where 13 of the planet's 17 major fisheries are in or near a similar state of collapse.

"Are we learning about the environment fast enough to overcome the impact we are having on it?" Lynch asked in a lecture a couple years ago. That question—our own human enigma looms large right now over Georges Bank and the world's other troubled fisheries. Perhaps part of the answer to it, too, thrums through the chips and wires that link 229 Cummings Hall to the rest of the world.

Hypothetical "source" population on Georges Bank (top), dispersed by weather and currents.

The Data Bank Every intersection on this finite element mesh represents a data collection point used in Lynch's mathematical models.

The Big Equation Dozens of differential equations boil down to this one: the formula for measuring the growth and decline of the cod, in raw numbers.

In the salad days of the fifties Georges Bank fishermen dumped their catch on the Boston fish market.

The Dartmouth Players Dan Lynch is just one of a small army of researchers using modeling to help solve the Georges Bank enigma. Below are current and past colleagues with Dartmouth degrees. Christopher Naimie, Ph.D. '95 Matthias lohnsen, Ph.D. '94 Justin Ip, Ph.D. '93 Elizabeth Wolff '92, MS '94 Monica Holboke, Ph.D. '98 Rebecca Sullivan '92, MS '94 Wendy Gentleman, Ph.D. candidate Curtis Thalken, MS '93

Journalist DAVID DOBBS wrote about Dartmouth's Hubbard Brook research study last January. He is currently writing a book about the Gulf of Maine fishery.