For a decade, marine biologists have been trying to unravel the subtle ways sonar may harm whales and dolphins, which hunt and travel using echolocation.

But experts peeling back the role of sound in the marine world are making surprising observations.

Alongside a boom in international shipping and deep-sea oil development, the ocean is growing ever-noisier and scientists are increasingly wary of sound's potential to impact sea life beyond just marine mammals.

Herring and cod appear to alter their swimming patterns in response to noise from ships. Schools of bluefin tuna scatter, with some diving and others rising to the surface. New studies suggest even small bumps in ocean noise may affect everything from damselfish and pollock to octopus and squid.

"The new research has been an eye-opener," said Jason Gedamke, who runs the US National Marine Fisheries Service ocean-acoustics programme.

The study of human-caused underwater noise pollution is in its infancy. For most creatures it's too soon to say how much is too much.

But the issue is capturing attention at high levels. Among the last acts of the top two science advisers in the George W Bush White House was a report recommending a decade-long research plan to grasp the "biologically significant effects" of marine noise.

And last year, leading experts on the sea's auditory environment compared the potential harm from ocean noise to Rachel Carson's 1962 plea to curb the use of toxic pesticides.

"Studies on the impact of pesticides on birds ... have curbed the prophesy of a 'silent spring'," they wrote. More noise research will provide "a better alternative to waiting to see what happens to fish in the dim future of a 'noisy spring"'.

The Bush administration was responding largely to gridlock from fights over sonar's impacts on marine mammals. But it recognised that our understanding of sound in the sea was changing rapidly.

In the last decade, beaked whales washed ashore with bleeding around the brain shortly after exposure to mid-frequency sonar. Researchers figured out that stranded bottlenose and rough-toothed dolphins often were nearly deaf. Dall's porpoises and killer whales were found to alter travel patterns during Navy exercises in Haro Strait, off the far north-west coast of the United States.

Environmentalists repeatedly sued the US Navy, but the precise science behind the impact on whales was often elusive.

Only recently did new research finally suggest, for example, that Navy sonar may mimic sounds produced by predatory killer whales. That may drive prey like beaked whales away from feeding areas and send them rocketing to the surface, giving them the equivalent of the bends.

But this acoustic war over mid-frequency sonar masked another emerging issue. At least 800 species of fish hear and produce sounds, either while fighting or competing for food or when courting or spawning. Some, like herring and shad, can detect ultrasound, which may be how they avoid hungry whales. Even sharks, considered poor listeners, follow sounds, perhaps when they resemble noises made by wounded prey.

The noise pollution emanating from shipping lanes has increased more than tenfold since the early 1960s. And while higher-frequency sonar may be harmful to animals nearby, the low-frequency groan from shipping and the deep-sea air guns used to build oil platforms and bridges can travel halfway around the world.

"Navy sonar got our attention but now we're looking at low-frequency noise and thinking, 'wow, this could be very important'," said Aran Mooney, a zoologist with the Woods Hole Oceanographic Institution in Massachusetts.

Marla Holt, a marine biologist at the Northwest Fisheries Science Center in Seattle, said: "We've opened up this whole new area of research and are looking at acoustic exposure in new ways."

Now, rather than the discreet, damaging event, "it's that consistent, ubiquitous exposure that may be the most concerning", she said.

Pumping up the volume of background noise has well-known physiological effects on humans. It can alter hormone levels and increase blood pressure, and prompt us to change our behaviour. Same too, with some fish, which boost the release of the stress hormone cortisol in response to white noise.

"The analogy I use, having grown up in New York City, is when you see someone jackhammering a sidewalk, do you go out of your way to walk around it or do you just walk right by," said Art Popper, a biology professor who runs the aquatic bioacoustics lab at the University of Maryland.

The impact of sound varies by species. Goldfish have excellent hearing. Salmon and trout don't. Humpback, fin whales, right whales and bowhead generally flee all types of noise. But how important a sign is that, really? Does it mean animals are driven from important feeding or mating grounds, or is it merely a nuisance?

A chief concern is the potential for "auditory masking". Killer whales, for example, have been shown to raise their voices when their group-specific calls are being drowned out by noise.

But is that as harmless as humans talking a little louder - or is it the equivalent of regularly screaming at the top range?

"Is there an energetic cost to all of this?" Holt asked.

Scientists don't really know.