A new technique for hail prediction by Boulder researchers
These days in Boulder, Colo. — at the National Center for Atmospheric Research, scientists are working on new technology which uses the same algorithm as facial recognition software in cell phones to predict hail storms better. It could determine hail-producing storms and the size of hailstones.
“This can help forecasters by helping them deal with a large amount of data they receive every single day,” said David John Gagne, a machine learning scientist at NCAR.
By setting up the facial recognition application in a cell phone, the software takes pictures of the face of a user to create a machine learning model.
“That model will look for patterns in your face to say, ‘this face is you and not someone else’s,’” Gagne said.
Following this, each time before granting access to a user to unlock the phone, it will take another picture for comparing the patterns together.
In this new technology, the same algorithm is used for weather patterns.
“Instead of giving it pictures of faces, you give it pictures of storms,” Gagne said. “We can use the same type of algorithm to look at large numbers of storms and differentiate the ones that produce large hail from the ones that produce small hail.”
It is expected that by this algorithm the place where a hail storm is to hit and largeness of the stones would be detected.
“Hail was a really untapped area for improving predictions,” he said.
Statistics show Colorado and Wyoming to be the hailstorm capital of the country. Colorado has hard-to-predict weather due to blocking the air masses and changing the direction of the winds by 14,000-foot wall the mountains create.
The uncertainty in the weather predictions would be reduced by this type of technology and it would be easier to warn the public of incoming storms and possible damages which helps them to take steps to protect themselves and their property faster.
In order to get to this long-awaited machine learning model, the first step was establishing a database. To do so, researchers had to upload information about several previous storms into the machine learning model.
Later on, to predict hail storms the model would search for similar patterns by using that data.
“What the machine learning models will do is look at slices of those storms, so it’s sort of like you’re looking at an MRI looking at slices of your body,” Gagne said.
Meanwhile, due to insufficient data, it is difficult to predict if a hailstone will be bigger than a golf ball and establish an accurate pattern.
This technology which Gagne has been working on its concept since 2015, can be used for prediction of hail storms which would occur between 30 minutes and one day in advance.
This technology helps meteorologists for preparing the public for severe storms and reducing the possible damages.
“Hail, in particular, has become a major concern for insurance companies because they have experienced more than $10 billion in losses per year every year since 2008,” Gagne said. “If this tool can help make better decisions, then you might see your lower insurance losses and then also hopefully lower insurance premiums.”
However, the researchers would like to make this technology applicable for predicting snow, lightning, tornadoes, hurricanes, flooding and more.
“These are really exciting times. Machine learning is just exploding in the both the scientific community and in the general public. I’m really excited to see where the research goes,” Gagne said.
The team is hoping to expand this technology more and shift it from experimental level to a fully operational model with the National Weather Service within the next year. Consequently, they’re hoping to present a more advanced national model within the next two to three years.