Not quite two years after federal scientists blasted water from four Glen Canyon Dam release tunnels to simulate floods of pre-dam days, those experts acknowledged Tuesday only limited success in replenishing Colorado River sandbars and fish habitat.

Although the 60-hour flush in 2008 took maximum advantage of the natural flooding on the Colorado tributaries, 90 percent of the sand that would replenish the river's ecology remains trapped behind the dam.

Theodore Melis, deputy chief of the U.S. Geological Survey's Grand Canyon Monitoring and Research Center in Flagstaff, Ariz., said the 2008 test adds to the evidence examined from controlled floods in 1998 and 2004.

"What we saw [in 2008] was a direct benefit," Melis said during a telephone news conference.

The flood of March 2008 reached a peak flow of 41,000 cubic feet per second and stirred up the gravel in the riverbed, Melis said, giving rainbow trout -- a non-native species -- a better shot at survival. The timing seemed also to reduce tamarisk seedling germination.

The high-flow test also pared the nuisance New Zealand mud snail population by about 80 percent, the USGS said, while midges and black flies (high-quality food items for fish) increased.

It was unclear how the flood affected habitat for the native humpback chub, an endangered species that has died off as the sand dwindled below the dam, completed in 1963.

Six months after the test, the USGS reported, new sandbars had been largely eroded by fluctuating dam flow operations, driven by electrical energy demand.

That will be the norm for the tests unless water releases from the dam even out, said Rick Johnson, Colorado River science director for the Grand Canyon Trust.

"The science has been around for a long time," Johnson said. "The solution is you run these flood flows frequently and between them, you stabilize the flows."

That would slow erosion and make the results of the controlled floods last longer. "The only thing missing," Johnson said, "is the political will to do it."

Melis estimated the cost of the 2008 flood at $3.5 million, in part, he said, from lost hydroelectric power revenues because the turbines were shut down for the test.

In 2008, federal officials said the power cost alone was $4 million.

National Park Service officials want regular, controlled floods to get more sand into the river, particularly after tributaries, including the Paria and Little Colorado rivers, have flooded naturally. The humpback chub's survival depends on the warm waters of the Little Colorado and the sandbars that keep the young fish from washing away.

The chub, once down to just 4,000 adults, has recovered about 1,000 of its numbers, giving encouragement to controlled-flood advocates.

Sand deposits from controlled floods also could help protect archaeological treasures, including 150 ancient sites and cliff dwellings, and the beaches river-runners camp on as they travel the 296 miles from Lees Ferry to Lake Mead.

In December, Interior Secretary Ken Salazar called for more controlled floods. "We must find a way," he said, "to protect one of the world's most treasured landscapes, the Grand Canyon, while meeting water and clean energy needs in the face of climate change."

The purpose of the Glen Canyon Dam, however, is to generate electricity. About 1.3 million customers now pay for the hydropower.

The U.S. Bureau of Reclamation has estimated moving a tenth of the sand trapped at the upstream end of Lake Powell would require a $100 million pipeline.