GRANTS PASS, Ore. — A federal review of the water bank that spent $7.6 million this year to buy irrigation water from Klamath Reclamation Project farmers to increase flows for salmon finds it is seriously drawing down the local water table and may not be able to help fish through extended dry periods.

In a report released yesterday, the U.S. Geological Survey (USGS) said a preliminary estimate has found that private wells pumping for the water bank have increased demand on the local groundwater eightfold, causing water levels in other wells to drop between 2 and 20 feet.

"If you were to continue pumping groundwater at that rate, eventually you would run into trouble," said Dennis Lynch, director of the USGS Oregon Water Science Center in Portland, which did the review. "Our thinking is groundwater is part of the solution. But there would have to be times when water levels were allowed to naturally recharge during wetter periods."

Christine Karas, deputy area manager for the Klamath Reclamation Project, said agencies need to look at ways besides the water bank to provide water for fish long term.

Lynn Long holds a sprig of wheat grown on the Klamath Reclamation Project. Long and other farmers are fighting a move to increase the rate for electricity to move irrigation water on the project to the same rate paid by other farmers around the state. Environmentalists hope increased electric rates will lead to less irrigation.

"In the meantime, the water bank is a great tool to help us," she said. "As far as a long-term sustainable solution, it's unlikely simply because of the high cost."

After shutting off irrigation water to farmers in 2001 to assure water in a drought for threatened and endangered fish, the U.S. Bureau of Reclamation started the water bank in 2003 to meet terms of the Endangered Species Act and responsibilities to Indian tribes that depend on fish.

This year, the bureau is paying $7.6 million for 100,000 acre feet of water devoted to endangered suckers in Upper Klamath Lake, the primary reservoir for the reclamation project, and threatened coho salmon in the Klamath River, the outflow of the lake and the project.

It is made up of 60,000 acre feet from paying farmers to turn their wells into irrigation canals, and 40,000 acre feet from paying farmers not to irrigate their land.

Lynch said it would be best not to have a water bank every year, so that groundwater depleted by the wells could recharge.

"It's kind of like having a bank account and only making withdrawals," he said. "At some point you run out of money."

The review also suggested going outside the irrigation project, as far away as the Shasta and Scott river valleys, for water to spread the impact of pumping wells.

The review said a better set of river-flow targets for wet, normal and dry years could be crafted by basing them on river flows going back to 1960, rather than just the 1990s. The current system suffers from large jumps in flows from wet to dry years, which would be smoothed out by using a longer timeframe.

Lynch said paying farmers not to irrigate can provide more water without threatening groundwater but does not offer water in the spring, when young salmon need it to migrate to the ocean. Water can't be spilled out of upper Klamath Lake in the spring, because it is needed to maintain sucker spawning habitat.

Building more reservoirs would help, but only if they were big enough to carry over water one year to another, Lynch said. One problem is that this reduces high winter flows in the river that some scientists think are needed to flush out parasites that kill salmon.

Farmers, environmentalists and salmon fishermen all found something encouraging in the review.

Bob Gasser of the Klamath Water Users Association liked the ideas of increasing reservoir storage and spreading the demand for the water bank outside the irrigation project.

"We understand we have to be part of the fix, but we can't be all the fix," Gasser said.

Glen Spain of the Pacific Coast Federation of Fishermen's Associations said the review underscored the idea that water in the basin has long been over-appropriated and that flow targets for the river need to be improved.

"The only ultimate solution here is to reduce demand, and retire water permits on a willing-seller basis to bring demand back in balance with actual supply," Spain said.

Steve Pedery of the Oregon Natural Resources Council said he hoped the bureau would follow up on the review's suggestion to consider restoring Lower Klamath Lake, drained for farmland, to increase water storage.