Machine learning, an area of artificial intelligence, provides greater opportunities to handle the vast amount of information available in a storm. This eliminates incorrect calculations four to six days before the hurricane strikes.
By applying machine learning to possible routes a storm can take in its advance, researchers at Penn State University were able to help meteorologists make better hurricane forecasts. Thus, communities at risk can be warned earlier.
“We investigate hurricane forecasts around the world and use statistics and machine learning to break down each forecast into possible scenarios,” Jenni Evans, professor of meteorology and atmospheric science, told phys.org .
Even with machine learning, however, as always with weather forecasts, there will be a margin of error.
– The models that are available are not perfect because you can not see every water molecule or every radiation from the sun, says Jenni Evans.
