Walking to the mouth of the Elwha River, Mike McHenry, a fisheries habitat biologist for the Lower Elwha Klallam, a tribe with a reservation on the lower river, wheels and holds up a piece of paper. "Before you even take a look at the river, look at this." A 1930s aerial photo shows a wide, branching river winding through forested islands before spilling over a beach that, McHenry says, was filled with shellfish.

"Flash forward to today." The river is now a single brown current, racing toward the strait. The dams lacked fish ladders, so salmon couldn't migrate upstream past the first five miles of the river, and they siphoned off the small boulders, dirt and huge logs that would have helped slow the current and create spawning beds. The river was left studded with large rocks, emptying onto a beach now too stony for the clams the tribe relied on.

McHenry's office computer is filled with historical photos, including a 1920s picture of a grinning Ernest Sampson Sr., a tribal member, with a chinook slung over his shoulder. The fish is so large that its tail touches the ground. By then, however, the river was already locked in its slow decline. Pink salmon populations crashed midcentury. The spring run chinook dwindled and the monster fish disappeared, even in stories. Sockeye were extirpated. Now, three of the river's species are listed under the Endangered Species Act: Puget Sound chinook, bull trout and Puget Sound steelhead.

To help the fish, the three-year-long dam removal has built-in "windows" when sediment flow will be halted so salmon can spawn in water not choked with dirt. McHenry and his technicians (including Sampson's grandson) are also building 50 log jams, stacking huge tree trunks and lashing them together with cables to trap sediment and create pools for salmon. Some of these will eventually form islands that provide shelter for mink and roosts for bald eagles, McHenry hopes.

In addition, as Lake Mills empties, Olympic National Park fisheries biologist Sam Brenkman and his crew have been catching bull trout so they can be helicoptered upstream out of harm's way. They marked about 30 with radio transmitters, so as the dams come down, biologists can track whether the fish are finding spots to persevere, competing with newcomers from downstream, or making their way to the ocean now that the barriers are gone. Brenkman and others snorkeled the entire length of the river several years ago, counting fish and charting habitat to gain a baseline for comparison when the dams come out. "We're setting the stage. No one's really done this kind of work before," he says. They'll do a follow-up survey in 2014, if funding permits.

The vanished reservoirs will leave behind bowls of mud at risk of being taken over by weeds. A native plant nursery will provide seedlings so botanists can rush in ahead of invasives like reed canarygrass and Scotch broom to plant common snowberry and Nootka rose.

The goal is not to recreate snapshots from long ago, according to Joshua Chenoweth, botanical restorationist for the park, but rather to provide flexibility in the face of an uncertain future: "There are so many variables, our concern isn't whether it turns into a Douglas-fir forest or a hemlock forest." But any kind of forest might be hard to grow. The exposed ground, in some cases, will be just sand and gravel. Few young trees will take root in it, even species like red alder, which are some of the first to appear at disturbed sites.

In the future, due to climate change, the river may be warmer and slower, fed more by rainstorms than snowmelt, attracting a different suite of species than were there a hundred years ago or requiring existing species to adopt new life histories. The ability to survive in the face of disturbance is termed "resilience," and it is one of the major goals of restoration projects like this one. Large numbers and genetic diversity both help a species persist.

Perhaps the most controversial element of the Elwha Fish Restoration Plan, put together by the park, the tribe, the state and NOAA, is the use of hatchery fish, particularly a non-native strain, the Chambers Creek steelhead, which matures more quickly than its native counterpart. Hatchery fish have a lower survival rate and carry disease, and may put the resilience of wild fish at risk. Restoration money funded a new hatchery for the tribe, which, together with a state facility, will raise chinook, coho, steelhead, chum and pink salmon to be released, some right at the hatcheries, some in the other reaches of the river, before, during, and perhaps even after dam removal.

Larry Ward, manager of the new tribal hatchery, says the hatcheries will increase stocks for tribal members who rely on fish for their livelihood. The Lower Elwha Klallam have already made concessions by agreeing to a five-year fishing moratorium on the river, he notes. Though scientists might consider the river a natural laboratory to study wild fish, he adds, it isn't: "There are not really stocks on the river that haven't been influenced by the hatchery at some point."

Some biologists think salmon will recolonize the river naturally and don't need help. Farther upriver, McHenry points out one of the river's side channels. Not far from the raging, muddy body of the Elwha, a clear stream cuts through the brush; small stones and gravel line its bed. It's good spawning ground. "If you're a fish in here," he says, picking up stones, "you got a good chance of surviving dam removal."

Those who want to track wild fish after the dams come down face obstacles besides hatchery interference. Even with the $325 million and all the studies under way, money for post-dam-removal monitoring is in short supply. So far, scientists and agencies have been patching together smaller grants year by year, hoping to study how the ecosystem responds to salmon recolonizing, how species interact, how the forest regenerates or doesn't regenerate. According to McHenry, "We've made great progress ... but it doesn't look like we're going to have the resources to carry it all the way through. My biggest fear is that we won't be able to answer some of the questions that society's going to want to know. ... That would be a failure of this project in my mind."