Is it raining or snowing? Here’s why that’s so hard for scientists to answer.
Scientists are using lasers, satellites and even sound to determine what is falling from the sky—but the best tool may still be the human eye.
From kids with hopes of sledding to adults wanting to avoid slippery roads, northern winters bring a lot of anticipation of snowfall. But the simple question “Is it snowing?” is more complicated than it appears.
On a remote mountaintop in Colorado, the Storm Peak Laboratory has a strange winter setup: scientists scurry around, placing more than 30 radar and sampling instruments in different parts of the mountain, on different frequencies, to get at a simple question: what, exactly, is falling from the sky? What are the sizes and shapes of snowflakes, and when is that precipitation actually rain?
People can look outside the window and judge snowfall very quickly, but it’s challenging for technology, even with advanced tools, to do the same—and that has big impacts. Cities need to know how much snow is falling—and how dense it is—in order to decide whether to gas up plows and deploy them. Accurate measurement is also important for highway managers who control road closures. Water managers need to know what is falling because winter snowpack becomes the water that often feeds agriculture in the months following, whereas rain typically washes downstream.
Snowfall is a question that has beguiled scientists for decades. "It should be a really simple problem,” says Keith Jennings, director of research at the University of Vermont’s Water Resources Institute. “Is it raining? Or is it snowing?” Part of the problem is that the temperature doesn’t determine what comes down. In a spread of nine degrees Fahrenheit, precipitation can be rain or snow, depending on what it experiences in the upper atmosphere.
With new technology and tools, researchers tackle this age-old issue every winter, and they are getting closer to models and techniques that accurately assess what’s coming down. Some of the best methods, it turns out, still may include rudimentary human involvement.
Buckets and bytes
Why not just put out a bucket and weigh what goes into it? Claire Pettersen, a researcher at the University of Michigan gets that question a lot. The problem, she says, is that any little breeze pushes out flakes—especially if they’re fluffy, light ones. “You could wave your hand in the air, and it’s going to move that snow all over as it’s falling,” she says.
Sale Extended - Save up to $20!
One of her recent studies used a video distrometer, an instrument with high-speed cameras, to capture images of falling precipitation from different angles. That data, combined with machine learning, showed that there are nine different types of precipitation between rain and snow, ranging from drizzle to heavy snowfall. Each has a unique pattern.
Pettersen says that combining methods, like she does in experiments at the Storm Peak Laboratory, may yield the most accurate readings of what is falling from the sky, from satellites in space to lasers on the ground, that can be broken up by flakes falling to planes that fly through storms to understand heavy snow processes. “The best case scenario is you have a bunch of different instruments at your fingertips,” she says.
There are also more experimental ways to measure precipitation. Another research group at the University of Vermont is using acoustic sensors to determine the sound characteristics of different types of precipitation.
Each research approach comes with challenges. In mountains, it’s hard to keep equipment running. And in places like New England, ice storms often knock out sensors and technology.
“You could have the best piece of equipment that you finally get into a mountain area where you want to get these measurements, and then you have a big ice storm come through, and then you've got to charge back up there, clean all the ice off the equipment – it can be quite the process,” says Jennings. The future may include more automated observations that don’t require human intervention, he adds.
Nothing like the human touch
Sometimes humans still reign supreme in discerning the weather. When the temperature is between 32 and 45 Fahrenheit, it’s hard for computer models to figure out what’s falling from the sky. Sometimes known as “wintry mix,” this slop can be anything between rain and snow.
Five years ago, Jennings and his colleagues launched a citizen science project called Mountain Rain or Snow to crowdsource visual reports of rain, snow, or mixed precipitation in mountain range areas through an app. They hoped human-submitted reports would improve models.
So far, they’ve captured nearly 100,000 observations from 1,700 observers around the world. When researchers tested three AI algorithms with the data from volunteers, along with historical weather data, the technological tools were barely better than traditional meteorological methods when temperatures were close to freezing.
"A few years after we've started this project, we still, and perhaps even more strongly now than ever, believe that the best way to monitor the type of precipitation where it's falling is with human observers,” says Jennings. “It’s quite exciting and maybe counterintuitive in our era of artificial intelligence, but people still play a very large role."
When snow is on the ground, people are still useful, as well. The basic technology for measuring snow is still remarkably simple – a ruler, a spatula, a rain gauge and a flat white board. More than 20,000 Americans help out by measuring precipitation through the Community Collaborative Rain, Hail and Snow Network., which was founded in 1998. The data they collect goes into a central database and is used by the National Weather Service, as well as federal efforts to monitor droughts and manage water resources.
As technology and models improve, researchers are chasing a climate model that is also changing. Mountains are some of the places hit hardest by climate change, and that often means snow becoming rain. It presents two challenges, Jennings says: one is that these areas have poor historical data, so it’s hard to know a baseline. And the second is that they are transitioning into a climate that is likely to be more rainy – which can wash away snowpack.
“It's really a big open question on what that snow to rain shift is going to look like over the next 50 to 100 years,” he says.
