For years, scientists and researchers pointed to nutrients from septic systems as a leading cause of the massive fish kills that repeatedly wiped out sculpins, rockfish, perch, sea stars and dozens of other marine creatures in Hood Canal.

But the most comprehensive review ever of existing research on Hood Canal has come to a different conclusion.

A new joint report by the federal Environmental Protection Agency and the state Department of Ecology determined the link between human activity and the hooked fjord's low-oxygen problems wasn't solid enough to warrant setting new strict pollution limits.

The two watchdog agencies — along with an exhaustive review by outside experts — concluded that human sources of nitrogen contributing to low-oxygen events in the mainstem of Hood Canal were "insignificant," while evidence linking humans to oxygen problems in the more troubled area near the canal's end in Lynch Cove "is not strong."

The overwhelming causes of fish kills, the agencies concluded, are the geography of the canal and ocean conditions.

The findings complicate efforts to resolve an array of problems facing Hood Canal and highlight the difficulty of fully understanding the complex variables that can harm marine life. The role of human development in the low-oxygen events had become accepted wisdom among many policymakers.

"Some earlier reports probably implied that human sources were a much larger contributor," said Tom Eaton, Washington state policy director for the EPA. "Our best assessment is that Lynch Cove is sensitive, but we think the overall impacts aren't as significant as previously thought."

No one is suggesting that the 60,000 people living along the canal don't cause serious trouble for sea life. Bacteria from septic drain fields, particularly along sensitive shorelines, often lead to shellfish bed closures and may harm other aquatic creatures.

But the report reveals a divide among marine scientists about what role humans play in altering a cascade of natural processes known to starve waters of oxygen, periodically killing scores of fish.

"A lot of uncertainty"

Some believe that people are still a significant source of the problem, but that our technology and monitoring capability aren't yet sensitive enough to prove it.

"I would be the first to admit that there is a lot of uncertainty," said Jan Newton, a University of Washington scientist who led earlier investigations into Hood Canal oxygen problems. "The area from the great bend to Lynch Cove is the area where we think human nutrient loading has the potential to change oxygen levels. Some numbers suggest we are at that threshold. But not all of the numbers do. It's an area we should be looking at more closely."

Scott Brewer, executive director of the Hood Canal Coordinating Council, which represents local counties trying to figure out what they should do to keep the canal clean and safe, agreed.

"To me it didn't negate the idea that there are human impacts there," he said. "I think there are."

But some colleagues, such as Michael Brett, a UW environmental-engineering professor who evaluated nitrogen runoff from vegetation and septic tanks into Hood Canal, said that while climate change or other human-caused impacts may play a role, the study should put to rest the link between housing development and fish kills.

"What this shows is that it's not clear that humans are affecting (Hood Canal) much at all," he said. "What we're seeing out there is a fairly natural phenomenon. There may not be much we can do to influence things."

That question — what can be done — is the underlying concern.

How fish kills happen

Scientists long have understood the basic causes of Hood Canal fish deaths. Parts of the canal are deep, but its entrance is shallow enough that water circulates poorly in and out. Water also circulates poorly from surface to bottom.

Plants and phytoplankton bloom during sunny periods and then die and decay, and nutrients get brought in by rivers. When wind and climate conditions are right, all that decay is significant enough to rob parts of the canal of most of its oxygen.

Fish kills have hammered Hood Canal off and on at least as far back as the 1920s, if not earlier. A 2008 study of core samples found that oxygen levels in much of the canal were actually lower before 1900 than they were for most of the 20th century.

But in 2002, 2003, 2004, 2006 and 2010, periods of low oxygen were so extreme that fish died in massive numbers. In most of those years, thousands of fish from more than two dozen species were killed. Underwater photographs captured wolf eels seeming to pant and rockfish by the hundreds gathering near the water's surface, where oxygen was more plentiful. Dead lingcod washed ashore.

UW fisheries professor Tim Essington has shown that crabs try to move away from these low-oxygen events. He's in the middle of a research project trying to gauge whether forage fish like herring starve or grow weaker because the tiny zooplankton they feed on hide at the bottom during these periods of low oxygen.

While no one formally tracked fish kills through history, the events of the last decade unnerved many scientists.

"In my experience in Puget Sound, which began in the late 1970s, what we experienced in the early 2000s was unprecedented," said former Washington Department of Fish and Wildlife research scientist Wayne Palsson. "The reason they got the attention they got is that it wasn't a really common occurrence."

Essington, too, said the intensity and size of low-oxygen events appears greater now than in recent decades.

Natural, human sources

When several studies, including by both Newton and Brett, showed significant nutrient loading from human waste found its way to Hood Canal, that seemed to offer an explanation.

But there was debate about scale. Even though hundreds of tons of nutrients washed into the canal, that remained a fraction of the nitrogen coming from natural sources. The chief question remained: Were human sources enough to push the system over the edge?

"The biggest source is the Pacific Ocean," said Mindy Roberts, who led the new review for Ecology. But in Lynch Cove, in particular, "some evidence is showing humans are contributing. Other evidence is saying no."

The state and federal governments' interest was in trying to decide whether to regulate the flow of nutrients into Hood Canal. But the evidence compiled from many studies — from UW, the National Oceanic and Atmospheric Administration, the U.S. Geological Survey and others — wasn't strong enough to support that.

"There are certainly other reasons to deal with old or failing septic systems," said Brett. "But this shows you can't predicate that on the assumption that it will help with fish kills."

Craig Welch: 206-464-2093 or cwelch@seattletimes.com. On Twitter @craigawelch.