In pools that teemed with hundreds of coho salmon just a few decades ago, researchers are now searching for signs of migrating fish in hopes of someday restoring the debilitated fishery of Lagunitas/Papermill Creek.

Historic levels of winter coho runs are estimated to have been between 3,000 and 5,000 fish annually. Nearly 200 were once counted within the five-mile tributary known as Devil's Gulch Creek alone. Oldtimers remember catching coho the size of a grown man's leg as the salmon swam upstream to spawn.

Now the tallies have dropped into the hundreds. Park Service biologists have spotted about 50 coho salmon in Devil's Gulch Creek so far this year, a meager count considering West Marin is home to 10 percent of this threatened species.

If enough genetic data can be obtained from the local population, however, scientists may be able to determine if the coho indigenous to West Marin are a unique stock that can never be replaced anywhere else - or perhaps a seed stock that can help replenish larger creek systems like those in the Russian River watershed.

As the salmon migrate upstream this season, biologists from Marin Municipal Water District, Golden Gate National Recreation Area, and local volunteers are counting fish and clipping fins (for genetic analysis) from any carcasses they find.

The salmon's rigid, three-year life cycle makes it difficult to survive a changing environment. Coho spend a year and a half in freshwater streams, then a year and a half in the Pacific Ocean before returning to their home stream to spawn and die.

In other words, a single creek's population must perpetuate itself every three years or it may cease to exist.

If the Lagunitas/Papermill Creek stock of coho salmon are indeed a unique subspecies, then they have been particularly devastated over the last 30 years, said Bob Chamberlain, a Trout Unlimited volunteer who studies the local salmon population.

"In the last 10 generations, this creek's gone from the largest commercial fishery [for coho salmon] to none," he said at his creekside home in Lagunitas.

Chamberlain said genetic testing is necessary to determine if the National Marine Fisheries Service's "Evolutionary Significant Units" of salmon along the coast are valid designations. Marin County coho are considered part of the Central Coast ESU, along with those in Mendocino, Sonoma, San Mateo, and Santa Cruz counties.

Now in its third year, the genetic-testing program has almost finished studying a complete generation of local coho salmon. At least two generations' worth of data will be needed to make any conclusions about the specificity of the local population, Chamberlain said.

The genetic information compiled by the Bodega Marine Laboratory will be shared by water district and park biologists, both of whom are working to restore salmon and steelhead habitat on the creek.

The creek has been dammed and diverted extensively over the last century-and-a-half by loggers, ranchers, and water district engineers.

In comparison, restoration work has been minimal. Under orders from the State Water Resources Control Board, MMWD has just begun a 10-year, $1.3 million project to reduce sedimentation of the creekbed and to enhance vegetation growing along the banks.

The project was required two years ago as mitigation for raising Peters Dam at Kent Reservoir in 1982, thus reducing the flow of water to the fishery downstream. One of the project's tasks already completed has been to control heavy erosion on a bare hillside beside the dam.

Future tasks include the installation of gravel traps to prevent fine sediment from filling in the streambed and placing woody debris in the channel to create pools where young salmon can grow during dry summer months.

A similar program is being undertaken by park biologists, who are in the first of a five-year, $800,000 project to assess and improve coho habitat on Redwood, Olema, and Lagunitas/Papermill creeks.

As heavy rains signal the anadromous fish to migrate upstream and complete their life cycle, biologists are counting on their powerful instincts and resiliency to see them through their perilous journey.

Somehow the fish have survived all the changes to the creek since Samuel P. Taylor built his papermill here over a century ago, said GGNRA biologist Ron Smith, who noted it will take a long time for them to recover to historic levels.

"We've barely gotten started," Smith said. "The course of restoration doesn't happen overnight....Over the years these fish have been studied to death. It's time to do something about it."