And don't overlook the strays. A small percentage of the migrating salmon don't return to the place they were born. Instead, they appear in other spawning sites, breeding with other distinct populations or re-colonizing creeks whose runs have been wiped out.

Other ecosystems operate with similar fundamentals derived from complexity. Wyoming deer and elk, for instance, ride a "Green Wave," roaming to feed where grasses and other forage are best at certain times of year, according to Armstrong, who's now based at the University of Wyoming.

Sea turtles, Nel informed me, are born on South African beaches and then swim through the waters of at least nine nations, navigating through threats like fishing nets. The turtles feast on jellyfish, but selectively, just nipping off the tentacles, much the way bears and gulls high-grade the best parts of salmon. Then the turtles return to lay eggs on the beaches where they began, following subtle cues including magnetic fields and smells –– "almost the same blueprint" and "ecological infrastructure" as the salmon, Nel said.

Any change that reduces complexity can threaten an ecosystem. Fragment the land here, with new roads and fences and other obstacles, and the bears can't roam freely enough to hit the brief peak runs on each creek, river segment and lake. The same goes for the water. In our walk along Lynx Creek, Schindler told me, "If a road went along here, it would constrain the stream and turn it into a rain gutter." The complexity of many varying temperatures and velocities in this single natural creek would be reduced, resulting in fewer options for salmon, and likely fewer salmon for predators, fewer nutrients derived from decaying salmon, and so on.

"An industry might say there's little or no impact from degrading or eliminating just one small salmon run, a few hundred salmon at one spot," Schindler said, "but it even reduces the complexity of time." Meaning, in such a strongly seasonal environment, if you subtract a few days from the peak run in this creek, you leave predators noticeably less time to fatten up enough to get through the rest of the year. Repeat that on enough creeks, and the bear population here would be winnowed down to the remnants in the Lower 48.

On another creek, Armstrong discovered one of the small complexities that would be easy to erase without noticing. He placed PIT tags (metal pins that can be detected by antennae) in juvenile coho salmon that spend most of their time in the headwaters, which on that creek are warm due to meanders and beaver ponds. Turns out, the tiny coho dash down to the creek's lower reaches to gobble the eggs of sockeye spawning near cold groundwater springs. Then they dash back up to the headwaters, because they need warm water to digest their meal. If a culvert or any other manmade obstacle is installed on a creek like that, it would cut off that feeding pattern – potentially causing a reduction in coho salmon, with who knows what ripple effects.

These are not hypotheticals. Scientists, fishermen and conservationists are alarmed about a proposal to construct the Pebble Mine on a high divide between two nearby watersheds, the Kvichak River/Iliamna Lake complex and the Nushagak River (for more info on the Pebble Mine, check the sidebar and the infographic/map). The rock bodies contain copper, gold and molybdenum, and are very porous and high in sulfides, so any runoff would be extremely acidic. And if the Pebble Mine is developed, more than a dozen other proposed mines would follow, Schindler said, using the roads and infrastructure built for Pebble.

Schindler is not an alarmist. He thinks that environmentalists often exaggerate when they talk about "fragile" ecosystems being threatened. Ecosystems in general are "resilient," he told me, in that their organisms and plants and inter-relationships can adjust to insults, "as long as we don't pave them over." But he's concluded that large-scale mining like that proposed for Pebble is incompatible with healthy watersheds and the fisheries and wildlife they support.

That's why Schindler and a few other scientists traveled to Washington, D.C., in early May 2013, as the federal Environmental Protection Agency revised its assessment of the Pebble Mine's potential impacts on the watershed. He met with EPA and congressional staff, urging the agency to scrupulously analyze the impacts. The EPA is under intense political pressure from congressional Republicans who are skeptical about a great deal of its science and heavily influenced by industry lobbyists. "The mining industry began promoting the Pebble Mine by saying a few years ago there will be no impacts," Schindler told me. "Now they admit there will be impacts, but they say they can mitigate it with hatcheries and buffering the impacted streams with limestone, and opening up new streams. In some cases they are talking about knocking down beaver dams. In other cases it is really not clear what they are talking about – possibly diverting water from non-salmon streams or digging spawning channels. It is all extremely vague. They have a list of consultants a mile long lining up to tell them what they want to hear."

The mine's backers, of course, tout its many human benefits – jobs, economic multipliers, more civilization for highly rural communities. But in any "manipulated landscapes," Schindler said, "we make them simpler by getting rid of variation we don't want or think we don't need." And then, when the ecosystem is damaged, "restoration" efforts are very difficult or impossible, because the complexity has been so reduced. "We also need to think about this complexity," he added, "when we talk about 'restoration' in the Lower 48."

Restoration has become a buzzword in the Lower 48. We dig up toxic mine sediments from riverbanks and haul them away, re-engineer meanders and pools on channelized streams, yank out culverts, install screens on irrigation diversions to block fish from straying into farmfields. On the land, we pull down fences or make them "wildlife-friendly" by tuning the arrangement of wires. We replant native vegetation where it was wiped out, tune up forests and wetlands that were mismanaged, bring in a toupee of topsoil to cover mine tailings and grow plants on roads carved for industry. In Alaska, I learned that we don't even understand the full extent of what was lost, and we probably never will.