Updated

As government engineers work to keep the nation's fourth-busiest seaport from losing its competitive edge, they are also planning what amounts to a massive science project to ensure fish in the harbor can still breathe.

When the Savannah harbor is deepened to allow for supersized cargo ships, the U.S. Army Corps of Engineers wants to install a dozen machines that function like bubblers in a home aquarium to compensate for an expected drop in dissolved oxygen. The 20-foot-tall steel cones suck water from the river, swirl it with oxygen from a generator until the bubbles break down and then pump it back.

Buying and installing the machines costs a hefty $70 million, plus yearly operating costs of $1.2 million. And the manufacturer says they've never been used for a project of this scale.

The oxygen machines are a piece of the $653 million proposal to dredge the Savannah River shipping channel to the Port of Savannah, a project officials hope will win final approval later this year. East Coast seaports are scrambling to deepen their harbors to accommodate supersized cargo ships expected through an expanded Panama Canal in 2014.

Deeper water will mean less oxygen toward the river bottom for bacteria, worms, shrimp, crabs and fish. But some scientists aren't sure the machines will be able to boost low oxygen levels as planned along 27 miles of the river.

"It's like putting the river on a respirator," said Chris DeScherer, an attorney for the Southern Environmental Law Center, which has sued on behalf of environmentalists who say the project would do irreparable harm.

The Army Corps of Engineers said in its final report on the harbor deepening this month that the Savannah River would lose relatively little dissolved oxygen overall. Still, it would dip below minimum standards set by Georgia and South Carolina, which share the river.

Pumping oxygen into water has been done since the 1960s, though usually in smaller bodies such as reservoirs and lakes, said Alexander Horne, an ecological engineer who has consulted on more than 20 projects to replenish oxygen in stressed waterways.

"It's like giving these systems an extra breath," said Horne, a professor emeritus at the University of California, Berkeley. "If I were a fish, I would vote for it hands down."

Rivers tend to be shallower and more turbulent, making them better at mixing oxygen on their own. But some still need a mechanical boost. A 12-mile stretch of the San Joaquin River in northern California installed an oxygen injection station in 2007. The Thames River in London uses barges equipped with bubblers.

Rivers naturally take in oxygen from the air, and their flows help mix it down below the surface. The deeper the water, the harder it is to push oxygen to the bottom. That's especially true in the Savannah harbor, where the land is flat and motion is slowed by pushback from the ocean tides.

Oxygen levels in the harbor, which stretches 38 miles from the Atlantic Ocean to the port upstream from downtown Savannah, are already stressed by decades of dredging. The river was deepened five times between 1912 and 1994, nearly doubling its depth from 21.5 feet to 42 feet.

Now the Corps and the Georgia Ports Authority want to scoop another 5 feet from the river bottom.

Standards imposed by Georgia and South Carolina say the river at minimum should have 4 milligrams of oxygen per liter of water. The Army Corps says the harbor's oxygen level tends to stay above 6 milligrams in the winter, but in places can sink to 3 milligrams per liter during the hot summers.

And they would decline further — by an estimated 3/10 of a milligram per liter or less — with an additional 5 feet of dredging.

Various machines using underwater hoses and pumps have been used to replenish oxygen in waterways. In Savannah, the Corps wants to use a variation called the Speece cone. Standing about 20 feet tall, the steel cones suck in water from the river and mix in oxygen from a generator until the bubbles break down and dissolve. The water then gets pumped back into the river.

The Corps plans to switch on the machines only during the hottest months — June through September. It's not a temporary fix.

"We expect to run them forever," said William Bailey, planning division chief for the Army Corps' Savannah District.

One species of particular concern in the Savannah River is the shortnose sturgeon, a primitive-looking fish with bony plates and a long snout. It's been listed as endangered since 1967. Blue crabs and striped bass also live in the harbor, but traffic from cargo ships makes it unpopular for fishing.

Shortnose sturgeons are bottom feeders. Deepening the Savannah harbor could shrink a portion of their habitat in the summer, the Corps concluded, unless oxygen is restored mechanically. The National Marine Fisheries Service, which monitors the endangered fish, last year called the proposal "a very risky operation with a high degree of uncertainty."

Steve Hatchel, president of the company that makes Speece cones, said he's confident they would work in the Savannah River, the company's largest project to date. However, while most projects use one to four machines, the Savannah harbor will need 12 cones to pump in 40,000 pounds of oxygen a day. State officials want them installed before dredging begins.

"We've worked on fairly large reservoirs and dams, but nothing that's been 27 miles long," said Hatchel of Indianapolis-based Eco Oxygen Technologies.

Two Speece cones were tested in the river in 2007. Results showed dissolved oxygen in the surrounding water rose by more than double the amount the Corps would need to restore.

But Paul Conrads, a South Carolina-based hydrologist with the U.S. Geological Survey, questioned that conclusion. Conrads compared the dissolved oxygen improvements recorded in the Savannah River with levels measured over the same period in the Cooper River at Charleston, S.C. He found oxygen levels rose in both rivers by the same amount — about 7/10 of a milligram per liter of water — though only one river had the machines.

Conrads concluded the natural tides likely caused most of the oxygen increase in both rivers.

"I'm not saying the Speece cones may not have an effect," Conrads said. "But with the data they collected, they could not quantify to me what the impact was."