Stories of unusually large ocean waves that appear without warning have loomed large in maritime folklore for centuries, killing sailors and leading scientists attempting to explain the phenomenon known as “rough waves.” It has confused me.
But a new study co-authored by oceanographers from the University of Victoria claims to have developed a machine learning model that can predict when and where these natural phenomena are likely to occur.
Researchers say the compound could save lives and protect the approximately 50,000 cargo ships that roam the planet at any given time.
Rough waves are elusive in nature, but are defined by deep troughs and steep walls that are at least twice the height of the surrounding ocean waves.
“People didn’t believe in them before,” says Johannes Gemmrich, a physical oceanographer at the University of Victoria and co-author of the study. Machine-guided discovery of real-world rogue wave models. “Sometimes a ship disappears and there’s no one left to tell the story.”
Things started to change in 1995, when 25-meter-high waves were recorded at the Draupner gas platform in the Norwegian North Sea.
“This was the first time we could actually measure it, but it still took time to convince people that it wasn’t a measurement error,” Gemrich said.
In the years that followed, two theories about rogue waves emerged. The first explanation states that rogue waves are created when one wave slowly extracts energy from another, and over time one very large wave results. His second theory, known as “linear superposition,” states that rogue waves occur when two or more wave systems temporarily align at their troughs and crests, producing one very large wave. Thing.
Gemrich believes he and his colleagues at the University of Copenhagen have proven that the latter theory is the most likely cause of the rogue wave.
“When two wave systems collide at sea, increasing the likelihood of a high crest followed by a deep trough, there is a risk of very large waves,” co-author Dion Hafner of the University of Copenhagen wrote. ing. “This is knowledge that has been around for 300 years and is now supported by data.”
Data for this study was collected from 158 buoys scattered around the world’s oceans collecting wave data around the clock. Researchers estimate that the data contains more than 1 billion wave measurements.
“We entered the dataset with 100,000 waves that can be defined as rogue waves,” Gemmrich says. “This equates to about one monster wave every day in random locations in the ocean.”
Armed with that data, the researchers used artificial intelligence to comb through the oceanographic variables that correlate with the appearance of rough waves, ultimately arriving at a recipe of sorts for monster waves.
The study authors are now focused on using algorithms to predict where and when conditions are highest for rogue waves to occur.
Gemrich said he plans to apply the algorithm next year to the North Pacific off the coast of British Columbia, where he hopes to be able to provide real-time red wave risk assessments for marine weather forecasting for commercial shipping lines and other mariners.
“The end result is that we have a forecast for the North Pacific Ocean that shows wave heights of about 6 meters and a very high probability of severe waves of 12 meters in height within two days,” Gemrich said. To tell.
“So if you want to ship something across the North Pacific and you think there’s a high chance of fraud off the coast of Alaska, you can choose a more southerly route.”
