Too Big to Fail?

June 5, 2014

North America is filled with lakes with outsized personalities: Utah’s Great Salt Lake, Nevada’s Lake Mead, formed by the construction of the Hoover Dam, Louisiana’s levee-protected Lake Ponchartrain.

Then there are the North American Laurentian Great Lakes. Individually, they are this country’s five largest lakes in surface area. They each have a distinct signature: Erie, for instance, host to the Rust Belt cities of Cleveland and Toledo, is shallow and vulnerable to pollution such as nutrient contamination. Superior is vast, deep, and apparently pristine, but subject to the effects of climate perturbations.

But the five Great Lakes are also joined together in an almost unthinkably massive hydrological system. Superior and Michigan flow into Huron, which flows into Erie, and then all of that water travels from Erie into the Niagara Falls and on to Lake Ontario.

With a coastline of more than 10,000 miles and the largest surface of unfrozen, fresh water on Earth, the five lakes are population and economic hubs—centers of industrial manufacturing and shipping, providing 1.5 million jobs and $62 billion in wages in the United States every year. Commercial fishing pumps an additional $1 billion into the regional economy each year, and recreational fishing even more. More than 37 million people live in the Great Lakes Basin.

You might say that the Great Lakes are too big to fail. Yet, despite years of management with some notable improvements, the lakes are still increasingly subject to environmental insults such as nutrient pollution, the introduction of invasive species from overseas shipping traffic, and fluctuations in water levels related to atmospheric oscillations and climate change. Recently, the water supply to 400,000 people in Toledo was abruptly shut off for several days because of a toxic algal bloom.

Two Duke graduates, Craig Stow PhD’92 and Andrew Gronewold PhD’08, are deeply involved in research on the most pressing environmental problems facing the Great Lakes. Stow and Gronewold were both trained in Kenneth Reckhow’s lab at the Nicholas School and also mentored by Robert Wolpert of the Nicholas School and the Department of Statistics. Now they are research scientists at the Great Lakes Environmental Research Laboratory (GLERL) in Ann Arbor, Michigan.

The lab is an outpost of the National Oceanographic and Atmospheric Administration. In fact, Stow and Gronewold say, the Great Lakes are more like oceans than lakes or rivers. “You can stand on 400-foot-high sand bluffs at Sleeping Bear Dunes National Seashore,” says Gronewold, “And you can’t even see the other shoreline of Lake Michigan. It’s like an ocean. There’s a real sense of awe.”

Even the equipment required to monitor the Great Lakes—large research boats and the type of sensors used to study deep oceans—is similar to what oceangoing researchers would employ.

It is the water level question that the two Duke graduates have focused on recently as part of the research team at GLERL. While residents and scientists on the saltwater coasts have been bracing for rising sea levels associated with climate change, the Great Lakes region has experienced low water levels and receding shorelines since 1998, the year of the most powerful ever El Niño Southern Oscillation on record. El Niños occur every few years when unusually cold water offshore of Peru and Chile drives climatic anomalies throughout a broad region, including increasing storms and rainfall in much of the United States.

Water levels on the Great Lakes typically fluctuate on both a seasonal and decadal timescale. Levels are relatively low in the winter, rise in the spring, and begin decreasing in the late summer. These shifts are associated with seasonal patterns in overlake precipitation and evaporation. There is also a larger 30-year pattern of fluctuation whose causes are not as precisely understood.

What is understood is that these long-term patterns changed after 1998. Lake levels dropped and didn’t rebound, eventually hitting record monthly lows. A New York Times piece on the problem described summer residents with lake views who now saw their docks as distant islands in a desiccated mudflat and marinas that were forced to dredge or lose clients. Shipping has been impacted, because deeper water is needed for navigation.

The questions for Gronewold, Stow, and their research colleagues were whether this shift was caused by the 1998 El Niño, whether the lake levels would rebound, or whether this was the “new normal,” part of a broader, longterm shift associated with climate change.

We often associate climate change with sea level rise, but Drew Gronewold explains how the opposite could occur in the Great Lakes. “We have a potential increase in precipitation and a potential increase in temperature associated with climate change,” he says. Higher rainfall would lead to higher water, but higher temperatures could lead to increased evaporation loss and therefore lower water.

He asks, “Which one might win out?”

On the other hand, it is possible that warmer winters, higher evaporation and lower water levels of the last 15 years were all artifacts of the powerful 1998 El Niño, and that a different climate pattern could return the lake levels to their historical averages.

The two Nicholas School grads and their colleagues began crunching data and publishing their findings on the water level problem several years ago. As they wrote in Science, “Water resource management planning decisions . . . hinge critically on determining the extent to which the water-level drop in the late 1990s was a state shift resulting from a strong climate perturbation; part of a progressive decline resulting from global climate change; or a consequence of engineering modifications to the hydrologic system, including historical channel dredging and the regulation of Lake Superior outflows.”

Then came the winter of 2013–14. While most of us would be happy to forget the Polar Vortex, the scientists at GLERL immediately speculated on whether the cold winter would be enough to jolt the water levels back to normal.

How would that work? As their research team reported in an article in EOS published late this summer, cold water and cold air temperatures would mean less evaporation, in theory. And the water was definitely cold in the spring, so cold that ice cover outstayed its usual duration. [See photo, page 28, showing ice cover in March.] At the end of April, Lake Superior was still 50 percent covered in ice, Huron 20 percent and Michigan 10 percent. “It has been 15 years since we’ve had any appreciable ice cover,” says Stow.

Would the late ice keep the water temperatures cold throughout the summer and fall, when the evaporation pulse normally kicks in?

When Dukenvironment talked to Stow and Gronewold in late July, they had both recently returned from family lakeshore vacations. “I can tell you, the water is still cold,” says Gronewold. The cold water may have disrupted swimming and other family fun, but it appears to have also disrupted the recent hydrological regime, possibly by keeping traditional late summer and early fall evaporation patterns from setting in. Water levels were just shy of normal for late summer, and water temperatures were distinctly below normal. The scientists were cautiously optimistic that the lakes wouldn’t recede dramatically this fall.

Whatever happens, both scientists feel that it will be years if not decades before we can fully understand the unusual hydrological events of the past 15 years. They cite the ecologist John Magnuson, emeritus professor, University of Wisconsin-Madison, who defends longterm research that measures the gradual and sometimes undetectable changes that take place on decadal timescales.

Magnuson writes, “Because we are unable to directly sense slow changes and because we are even more limited in our abilities to interpret their cause-andeffect relations, processes over decades are hidden and reside in what I call ‘the invisible present.’”

Stow has worked at GLERL since 2006 and Gronewold since 2010. They describe it as research-focused, like a small academic department. It is staffed with PhDs, many of them jointly appointed at the University of Michigan, as well as postdoctoral associates, and it offers opportunities for students from elementary to graduate levels. Several other Nicholas School graduates have worked at the lab or collaborated with Stow and Gronewold, including YoonKung Cha PhD’11, Song S. Qian PhD’95, and Ibrahim Alameddine PhD’11.

Gronewold’s primary focus at GLERL has been the water level research, largely from the perspective of computer modeling. He has been part of a team that developed an interactive software tool that allows the public to go online and look at various predictions about Great Lakes hydrological and meteorological conditions. It is being used by shippers, hydropower operators, owners of docks and marinas, and property owners interested in current and future conditions.

Stow also is involved in research on nutrient ecology and its effect on water quality. New invasive species of mussels that have been introduced in recent years are changing how the lakes process and cycle nutrients like phosphorus, he says, and he is involved in evaluating targets for that nutrient that were established in the 1970s in the Great Lakes Water Quality Agreement, a pact involving both the United States and Canada. (The fact that the Great Lakes are bordered by eight U.S. states, two Canadian provinces and several tribal entities makes policymaking for the entire basin especially challenging.)

If their research seems closer to the policy interface than that of other PhD scientists, that’s no coincidence. Because they are researchers in a government agency, Stow says, their research “has to be useful for public decisionmaking.” Shippers need to know when the water is deep enough for navigation; homeowners and businesses need confidence that there will be safe water for drinking, cooking, and washing. People who manage parklands and operate commercial recreational areas need to know that it is safe to swim or fish in the lakes’ waters.

All of these constituents look to GLERL for science-based information on which to base decisions. Stow says, “There is strong support in the Great Lakes basin for what we do.”

Their interest in and ability to interact with decisionmakers and stakeholders also is part of the DNA of the Reckhow lab at Duke. “I have always been interested in how science informs decisionmaking,” says Kenneth Reckhow, professor emeritus of water resources at the Nicholas School. “Virtually all of my PhD students got a taste of that, and most have continued in this vein. Both Craig and Drew are interested in how the science can be used and interpreted. They both have always had one eye on how the results of their research are going to inform decisionmaking.”

For the near term, the water-level work continues to capture their imaginations. “We may have just reached the end of a 15-year regime affecting this massive hydrological system,” says Gronewold. “But we won’t really know for sure for another 15 years.” Their goal is to continue “peeling back the deeper layers of the onion.” One layer is clouds. “We want to start looking at the extent to which regional cloud cover may have played a role in the water levels,” Stow says. “A lack of cloud cover may have kept the lakes warm for the last 15 years. We can look at data about this from shoreline stations, but there’s not a lot of data collection overlake. So we’ll look at satellite imagery for a statistical difference in cloud cover. And if we find that, we need to know what would drive that?”

“No one has looked at this before,” he says.

Lisa M. Dellwo is a writer specializing in nature, environment, science, and foodways, based in Down East Maine. Her most recent article for Dukenvironment was a profile of Matt Redmann MEM’11, an environmental manager at Harley- Davidson Motor Company (Spring 2013).