While working on a research sailboat gliding¹ over glassy seas in the Pacific Ocean, oceanographer Giora Proskurowski noticed something new: The water was littered with confetti（五彩纸屑） -size pieces of plastic debris², until the moment the wind picked up and most of the particles disappeared. After taking samples of water at a depth of 16 feet (5 meters), Proskurowski, a researcher at the University of Washington, discovered that wind was pushing the lightweight plastic particles below the surface. That meant that decades of research into how much plastic litters the ocean, conducted by skimming only the surface, may in some cases vastly underestimate the true amount of plastic debris in the oceans, Proskurowski said.

Reporting in the journal Geophysical Research Letters this month, Proskurowski and co-lead author Tobias Kukulka, University of Delaware, said that data collected from just the surface of the water commonly underestimates the total amount of plastic in the water by an average factor of 2.5. In high winds the volume of plastic could be underestimated by a factor of 27.

"That really puts a lot of error into the compilation³（编辑） of the data set," Proskurowski said. The paper also detailed⁴ a new model that researchers and environmental groups can use to collect more accurate data in the future.

Plastic waste in the oceans is a concern because of the impact it might have on the environment. For instance, when fish ingest the plastics, it may degrade their liver functions. In addition, the particles make nice homes for bacteria and algae⁵（海藻） , which are then transported along with the particles into different regions of the ocean where they may be invasive and cause problems.

Proskurowski gathered data on a 2010 North Atlantic expedition where he and his team collected samples at the surface, plus an additional three or four depths down as far as 100 feet.

"Almost every tow we did contained plastic regardless of the depth," he said.

By combining the data with wind measurements, Proskurowski and his co-authors developed a simplified mathematical model that could potentially be used to match historical weather data, collected by satellite, with previous surface sampling to more accurately⁶ estimate the amount of plastic in the oceans.