The mad scramble by B.C. fishermen to get in on a sensational Fraser River salmon run is great news for them, but a head-scratcher for the scientists that apparently didn't see it coming.

Last year, the sockeye stocks were so depleted the federal government called an inquiry.

This year, fishermen will scoop up what's believed to be the largest run in nearly a century, at about 25 million, raising questions about how much scientists know -- or don't know -- about the iconic fish's journey.

"It's easy to say the (Fisheries Department) can't count fish," said Simon Fraser University fish biologist John Reynolds, noting criticism levelled at the government.

However, Reynolds notes a study conducted by one of his colleagues examining fish predictions in Alaska and other parts of the United States shows equally imprecise results.

"Nobody else does it any better," he said.

The 32-hour sockeye fishery, which was set to end Thursday night, was the longest opening in several years.

Fishermen scrambled to haul in what they could, but are finding a different landscape than eras past when bounty was more plentiful.

Ice to pack the fish is in short supply, shops to tune up boats are a dime a dozen and at the mouth of the Fraser, in waters off the historic fishing village of Steveston, there's only one major fuelling station left.

With only two major canneries left in the province, the extra fish are a boon all the way up to Prince Rupert.

Workers there are back at the plant to process the excess fish, said Joy Thorkelson, northern representative of the United Fishermen and Allied Workers' Union.

She noted the average age of gillnetters is rising, with many in their 60s, as fresh blood steers clear of entering the industry.

More stable science would help, she said.

"It's important to have predictions because fishermen, of course, like everyone else, like to have an orderly life," Thorkelson said. "But fishermen have long been optimists and every season they think the fish are going to come back."

Estimates on how many salmon will return to the rivers each year is based on several methods, including mathematical models plugged with counts of how many adults spawned and how many juveniles left their lakes.

The Pacific Salmon Commission, an independent joint Canadian-U.S. body that advises the federal Fisheries Department, increased this season's stock assessments Tuesday after recent test catches suggested a banner year not seen since 1913.

Barry Rosenberger, area director for Oceans and Fisheries Canada, says the sensational numbers were unlikely, but not entirely unexpected.

The government's scientific forecasting document for this year's stocks, which presents a probability range, shows somewhere between a 10 to 20 per cent chance that 25 million fish would return. People who aren't enmeshed in the industry frequently only look to the 50-per-cent estimate: seven to 11 million this year.

"That's where the disconnect is," said Rosenberger, adding that while the long-term planning document is usually released in November, it was only released in June because they modified their approach to modelling.

Sockeye live for about four years, hatching from gravel nests in some 200 tributaries throughout the Fraser River basin.

They spend about a year in a lake before beginning their spring migration out to sea. Two years later, they return to spawn in the same grounds they were born.

"Technically, it's extremely difficult to follow juvenile fish once they reach the ocean -- you just lose them," said Reynolds, who holds the Tom Buell chair in salmon conservation.

Trying to determine numbers so far in advance is imperfect and highly difficult.

"It's like trying to forecast the weather, which does affect the fish, and everything else that can affect them many, many months in advance," Reynolds said.

Pre-season figures are far less accurate than in-season counts based on test fishing in places such as the Straight of Georgia and Juan de Fuca Straight, which are used by the government to set regulations around quotas and when the fishery opens.

"You can't fish off the forecast, you have to fish off the real in-season information," Rosenberger said, adding notices with updated predictions go out twice a week.

"People say, 'Well you got 30 million now, you should have been fishing weeks ago.' We didn't have 30 million weeks ago."

Another upgrade in the sockeye run size can be expected Friday, Rosenberger said.

But scientists are finally tweaking calculations based on new variables, Reynolds said.

They've now recognized that starting about 15 years ago, something big changed in the relationship between how many fish leave and how many return.

Temperature of the seas, food sources and predators -- all of which could be related to climate change -- as well as human intervention are factors that could play a role, he said.

Rosenberger agreed ongoing scientific advancement is necessary.

Only about one-tenth of the estimated 10 million sockeye that were supposed to appear in 2009 actually arrived, forcing the closure of commercial fisheries and aboriginal food fisheries for First Nations in the area.

The Cohen Commission is expected to tackle the question of why when it starts hearing evidence during the public inquiry next month.