WHY WOULD THE SUSQUEHANNA FLATS suddenly be full of bay grasses? Two years ago Michael Kemp was motoring across the northern end of Chesapeake Bay with a boatful of scientists and students, checking out reports that underwater grassbeds might be expanding along the famous shoals that sit at the mouth of the Susquehanna River. An ecologist with the University of Maryland Center for Environmental Science (UMCES) Horn Point Laboratory, Kemp has spent 35 years studying bay grasses and for most of those years those grasses have been declining throughout the Chesapeake.

Lanky and lightly bearded, he scanned the passing water as the 26-foot cruiser crossed the deepwater shipping channel along the eastern side of the Bay and then headed northwest towards the mouth of the Susquehanna. Moving up onto the Flats, Kemp began spotting bay grasses — a lot of bay grasses. Water stargrass and wild celery were there as well as redhead grass and coontail, and plenty of watermilfoil and hydrilla, two nonnatives that also live here now.

Kemp was motoring right into the middle of a mystery. These grasses weren't expected to be here, at least not this many. They began declining nearly 50 years ago, and then 40 years ago most of the grasses abruptly disappeared when Tropical Storm Agnes unleashed heavy and historic rains across the Chesapeake's huge watershed and sent floods of brown, silt-bearing water surging down all the Bay's great rivers.

The first victim of Agnes was the bay grassbed that Kemp was now motoring through. The Susquehanna Flats are the shallow-water delta at the mouth of the largest and longest river on the East Coast, a river that drains a watershed of 27,000 square miles, including parts of upstate New York and nearly half of Pennsylvania. Carrying runoff from so much rich farmland, the Susquehanna empties as much water and sediment into the Bay as all the other rivers in the estuary combined. In just one week in the summer of 1972, the floods of Agnes washed 20 years of sediment into the Chesapeake, much of it sediment long trapped upstream behind the big dam at Conowingo. Unleashed through roaring floodgates, all that sediment began burying bay grassbeds and oyster bars, altering the ecology of the estuary for decades.

Before the flood, the biological abundance on the Flats was legendary — especially among fishermen, hunters, and birdwatchers. The grassbeds were a gathering ground for shad and stripers, catfish and largemouth bass; they were a feeding ground for millions of ducks and geese. In one survey, biologists in the 1940s counted more than 1.2 million ducks in the Flats, including canvasbacks, redheads, and widgeon.

After the flood, the Flats went mostly bare for 25 years or more, with some grasses scattered around the edges, but only sparse patches of bay grasses dotting the shoals. The disappearance of these grasses on the Flats became an early signal of systemwide decline. When the grasses failed to come back after Agnes, when they kept dwindling throughout the estuary, they came to symbolize the plight of the Bay and the failure of Bay restoration efforts.

Gliding across the Flats in 2010, Kemp found himself cutting through bay grasses tall enough to reach up to the surface through six feet of water and thick enough to clog his propeller at low tide. In the late 1990s, the scattered patches of bay grass that survived Agnes began to slowly expand. Starting in 2005, the grass coverage suddenly quadrupled in only six years. As an ecologist, Kemp thought this expansion was explosive — and unexplained.

With the boat at anchor, Kemp clambered up on the roof of the cabin. The view from the top: bay grasses stretched in all directions as far as he could see. According to his maps and his quick math, he was looking at 25 square miles of healthy grassbed. He was, he admits, amazed. It was a once-upon-a-time vision: the Flats as they used to be.

He saw fishermen working the Flats for stripers, and ducks and geese working the beds for food. Perhaps it was time for ecologists to start working the big questions: why were the grasses coming back so fast? "This was an abrupt change," Kemp says now. "It was an incredible response to something — but we didn't know what."

Kemp had been surprised by bay grasses before. Like many ecologists he had studied examples of abrupt ecosystem change, but most of those changes were abrupt declines — not recoveries. Some 35 years ago he launched his career by teaming up with Walter Boynton at the UMCES Chesapeake Biological Lab and other scientists to organize a major study investigating the decline of bay grasses in the Chesapeake. The result was a set of unexpected findings about the causes of decline, findings that radically altered the scientific understanding of the Bay's ecology.

The bay-grass explosion on the Susquehanna Flats followed an eight-year window that brought no major storms or extreme river flows from 1995 to 2002. As a result, grassbeds were spreading rapidly across the Flats by 2001 and 2002. When Hurricane Isabel arrived in 2003, followed by Hurricane Ivan in 2004, the grassbeds were strong enough to take a one-two punch and get off the floor. Between 2005 and 2010, grass­bed coverage expanded fourfold. Figure by Cassie Gurbisz.

The major culprits, according to Kemp and Boynton and their colleagues, were not the usual suspects like toxics from industries or herbicides and pesticides from farm fields. Faced with a systemwide bay-grass decline, they focused on systemwide causes. Kemp and Boynton came to the Chesapeake as proteges of H.T. Odum, a founding pioneer of systems ecology and a proponent of a big-picture approach that focuses on how energy moves through biological communities, how communities organize into ecosystems, how ecosystems function, how they change. According to these newly minted systems ecologists, the key culprit in the bay-grass decline was the oversupply of nutrients that was flowing into and altering the Bay ecosystem. Washed into the Chesapeake, nitrogen and phosphorus were overfertilizing the growth of phytoplankton and other algae, creating enormous blooms that clouded the water, blocked sunlight from reaching bay grasses, and in their decay created dead zones along the bottom of the Bay.

Worse yet, nitrogen and phosphorus were coming into the ecosystem from everywhere: from the sewage plants of cities, suburbs, and towns; from the soil, manure, and fertilizer running off farmlands; they even came in from the atmosphere that carried the exhaust of hundreds of power plants and millions of automobiles. Cutting back on that inflow of nutrients became the central focus of a multistate campaign to restore the Chesapeake Bay.

Bay grasses had always played important roles in the ecology of the Chesapeake — and now they began to play a key role in the public's perception of ecosystem health. When people ask whether the Bay is getting better or worse, they want to know whether the summer dead zones are going away and whether the bay grassbeds are coming back.

Bay grasses had another surprise in store for Kemp. In 2010 the ecologist began investigating the bay-grass comeback by putting a graduate student to work. With funding from Maryland Sea Grant, he had Cassie Gurbisz start pulling together all the long-range data sets she could find on rainfall, river flow, temperature, salinity, water clarity, and 25 years of bay-grass surveys.

It's the kind of grunt work graduate students often get stuck with, but Gurbisz welcomed the opportunity. "We hear a lot of bad news about the environment, but this is a real example of good news," says Gurbisz, who once spent several years running field trips for the Chesapeake Bay Foundation trying to educate people about the problems facing the estuary. "It's cool," she says, "to be studying really good news and trying to figure out why it's happening and maybe help it happen in other places."

Ecologists Michael Kemp and Jeremy Testa gather bay-grass samples on the Susquehanna Flats (left). Graduate student Cassie Gurbisz hammers away at the support pipes that will hold a water sampling station, while research technician Debbie Hinkle lends a steadying hand (right). The station will suck up water every two hours and discharge it into 30 sampling bottles. Back at the UMCES Horn Point Lab, Gurbisz and Hinkle will analyze the samples for nutrients, chlorophyll, suspended sediments, and other water quality indicators. One platform will sit in the middle of a grassbed, another will sit outside the bed, giving research­ers data on how water quality affects grasses and how grasses affect water quality. Photographs by Debbie Hinkle (left) and Dale Booth (right).

A search for causes often begins with a search for correlations. When Kemp and Gurbisz started digging through the data, river flow emerged as the one factor constantly connected to changes in bay grasses (either recoveries or declines). Gurbisz calls it "the master variable." The remnant bay grasses left on the Flats after Agnes did poorly during wet years with high river flows, but they did much worse during years that brought big storms like Agnes (1972) or Eloise (1975), Isabel (2003) or Ivan (2004), or the huge, unnamed nor'easters of '93 and '96. During years of average river flow, on the other hand, these leftover grasses usually maintained themselves. They did better during dry years with low flow, and they did best of all during dry years with no major storms. Drought years may be bad for the farm economy, but they are usually good for the Bay's ecology.

If you want to write a formula for bringing back bay grasses, then delete storms from the equation. And delete them for several years in a row. Here's what Kemp and Gurbisz found in the data: the abrupt, fourfold expansion of grassbeds across the Susquehanna Flats followed an eight-year stretch that ran from 1995 to 2002 with no extreme flow events. That low-flow window, Kemp says, may have been enough to get a full-scale recovery started.

That's an important finding and something of a surprise. Since the floods of Agnes, other low-flow windows have come and gone without unveiling any major bay-grass recovery. Even in low flow years, after all, the Susquehanna was still draining sediment and nutrients off a huge watershed.

Yet water clarity, according to Kemp and Gurbisz, was improving on the Flats, suspended sediments and phosphorus were decreasing in the water column, and all these changes were clearly correlated with increases in bay grasses. Were all those sewage plant upgrades, all those new controls on farm runoff finally working? According to estimates, the flow of nutrients into the Chesapeake has been cut 20 percent in recent years. That's only about half the 40 percent nutrient reduction called for in the Chesapeake Bay cleanup plan, but it's now apparent that half a loaf can feed a recovery.

Half a cutback and a low-flow window let the grassbeds achieve liftoff, breaking out of a long-term equilibrium that featured low-density patches of plants scattered across the broad shoals of the Susquehanna Flats. According to Kemp's scenario, the clear water and calm weather helped grass patches spread quickly and link up into wider, denser beds. By the summer of 2011, the grassbeds on the Flats seemed to have reached a new, healthy, high-density equilibrium. But it was an equilibrium untested by big storms and high river flows.

In August of 2011, Tropical Storm Irene arrived. In September, Tropical Storm Lee also arrived, dumping even heavier rainfall across the Bay watershed. Lee unleashed the highest river flow in 15 years.

By early May of 2012, Kemp and his crew were back in their boat. Motoring across the Flats, they began seeing a lot of water that looked like chocolate milk. What they were not seeing was a lot of bay grass. The Flats held only some scattered stands of watermilfoil, but not much in the way of wild celery or water stargrass or redhead grass.

Scanning the water for a spot to take samples, Kemp decided the storm had destroyed the big grassbed and it was not coming back. Even though Lee had not been as huge as Agnes, history seemed to be repeating itself on the Susquehanna Flats.

When he came back in June, Kemp was able to find some grass along the eastern side of the flats, perhaps a remnant population, and his team started doing some biomass sampling. Better some data than no data. With some more hard data, Kemp could do some more theory building. Over several trips, he worked with Cassie Gurbisz and research technician Debbie Hinkle to set up water sampling stations inside and outside several grassbed patches. One of his research goals was a more detailed theory for how water quality could affect bay grasses — and how bay grasses could affect water quality.

As she worked, Gurbisz noticed a weird effect: there was more chocolate water in the middle of the grassbeds than outside. One of the ecological benefits of grassbeds is their ability to trap sediments and clear up cloudy water, but in the spring and early summer it was clear that sediments left over from Lee were being easily resuspended by wind and tidal action. As the summer progressed, however, more grasses began showing up and many of them were growing taller. As they began stretching to the surface, Gurbisz saw the water start to clear. And as the water got clearer, more grasses grew taller.

"When things start getting better, then 'positive feedback' will make them get better faster," says Michael Kemp. Photograph by Anne Gauzens.

Were the grasses growing because the water was clearer? According to one of Kemp's theories, the causality was running in two directions: the grasses were growing on the Flats because the water was clearer — and the water was clearer because the grasses were growing there. The plants themselves were reshaping their environment, making the Flats a better place for plants to grow. Kemp calls this process a "positive feedback effect," and he says it can be a strong force for recovery. "There is nothing subtle about the impact this bed has on the movement of water, the transport of sediments, the removal of nutrients, and a variety of other characteristics," he now says. "It is a dominant factor in that region." How dominant? At full strength, according to Kemp's calculations, the big grassbed in the Flats could absorb 5 percent of the total nitrogen entering the Upper Bay.

Here's where the going gets tricky. "Positive feedback" can sometimes have negative effects. "A simple fact about positive feedbacks: when things are bad," says Kemp, "the positive feedback makes them worse." When only a few bay grass plants are there, says Kemp, they can't help clear the water. Without clear water, new plants will not get started, and existing plants will disappear.

So why did bay grasses begin to return to the Flats? Part of the answer is a low-flow window with no large storms. Another part is clearer water, the result of environmental policies that are cutting down both land-based and airborne pollution. But an unnoticed piece of the answer is feedback, a natural ecosystem response unleashed by climate and clear water. There were now more bay grasses on the Flats to help clear up the water, that meant more clear water to help more bay grasses to grow, that meant more clear water, then more bay grasses, then more clear water, and on and on. "When things start getting better," says Kemp, "then positive feedback mechanisms will make them get better faster."

Systems ecologists talking about bay grasses can sometimes sound like physicists talking about the wave-particle paradoxes of quantum mechanics. "Positive feedbacks" that can also have negative effects is just one of the concepts in the intellectual toolbox of contemporary ecologists. As Kemp studies the rise and fall of bay grasses on the Flats, he also works with concepts like thresholds, equilibria, ecosystem regime shifts, and resilience, the ability of a biological system to withstand and recover from a major disturbance.

If bay grasses can survive on the Flats, they may yet provide a new narrative for understanding the plight and potential of Chesapeake Bay restoration efforts. Forty years ago, bad water and a big storm knocked down the famous grassbed, kicking off a feedback cycle that helped Bay ecology get worse faster. The disappearance of bay grasses on the Flats became an early warning signal that a Baywide decline was coming. Now bay grasses had reappeared, kicking off a new feedback cycle that could help the Bay's ecology get better faster. Perhaps an early alert that systemwide recoveries could be coming sooner than we expect?

Were the bay grasses back to stay? By the time his team wrapped up their 2012 field work on the Flats, Kemp had a better sense of the damage done by the barrage of big storms in 2011. There was deep scouring along the western side of the bed where the current was strongest, pushed there by the Coriolis effect created by the planet's rotation. "The storms really did have an impact," says Kemp.

But the bay grasses, at least those along the Susquehanna Flats, had one last surprise for Kemp: they passed the big storm test. He estimates that 60 percent of the grassbed survived the floods from tropical storms Irene and Lee, achieving lush green growth despite a cool spring and a short growing season. Good evidence for "resilience," one of his favorite concepts. "I would say that it's an amazing recovery," says Kemp. "If it weathered that storm, it is going to hang around for a while."