The death and destruction wrought by Hurricane Ida has yet again highlighted the shortcomings of weather forecasting and warning systems as climate change is making such storms more common and more powerful.
A Boston startup called Tomorrow.io is trying to solve that problem in an unusual way. It’s planning to use space technology to gather enough weather data to help scientists make more accurate, minute-to-minute predictions about killer storms and other catastrophic events, and ultimately create better climate models.
Already the six-year-old company has harnessed new data sources — such as disruptions in wireless network signals, video from street cameras, and sensors from Internet-connected cars — to provide much more rapid and precise hyperlocal weather forecasts.
It’s now aiming much higher: near-constant radar imaging of the entire globe, by launching a network of several dozen small satellites. Existing satellites feed pictures and infrared images of storms to weather forecasters, but radar data — which lets you see through clouds — comes only from ground stations and one massive NASA satellite that takes days to monitor the entire planet. Tomorrow.io’s smaller, cheaper satellites would fill in numerous gaps, particularly over the oceans.
Putting radar tech on a small satellite is no easy feat, but with Tomorrow.io’s system, its satellites would be able to measure how much water is in clouds or falling as rain, hail, or snow with much better precision than what is currently available, which is crucial for forecasting the impact of storms.
“We looked at what’s being done already, where the gaps are, and what we can do to help,” said chief strategy officer Rei Goffer in an interview at the company’s office in Bourne, where the space radar is being developed. “The biggest gap by far is around measuring precipitation, and the best instrument to measure it is radar.”
The young company already has experience providing highly detailed weather information to customers that include JetBlue, Uber, and the New England Patriots. For example, as the remains of Hurricane Ida were bearing down on the Northeast last week, Tomorrow.io’s customers had access to a dashboard and map showing problems such as flooding that would require sandbags or suspension of travel, down to the block level. Customers were able to see exactly where and when the storm would hit and how it would evolve over the next few days.
Tomorrow.io’s dashboard predicted Boston wouldn’t get hit as much as New York City, but warned that Logan Airport and sections of Storrow Drive and the Massachusetts Turnpike would be hardest hit by the heavy rains from Ida.
The company was also able to advise the US Open tennis tournament about the timing of rain storms as Ida approached New York. Other forecasts recommended canceling all matches that day, but the startup’s more precise information gave the green light for most matches to go on, accurately predicting that the rains would arrive later at night.
Tomorrow.io was started by Goffer and two other colleagues from the Israeli military in 2015, Shimon Elkabetz, and Itai Zlotnik, shortly after the trio arrived in Boston to attend business school: Goffer and Zlotnik at MIT, and Elkabetz at Harvard.
Hungry to get a business of their own started, the three friends brainstormed about using techniques Zlotnik had studied as an undergrad at Tel Aviv University to improve weather forecasts. Goffer and Elkabetz, former combat pilots, and Zlotnik, a veteran of Israeli special forces, had all encountered life-or-death situations in the field due to faulty weather forecasts.
One time, Goffer recalled, he had to land his plane through an unexpected fog bank at night with fuel running low. “We have just one shot at landing before we’re out of fuel,” he said. (He made it.)
Though still young, the company has attracted $185 million in backing and grown to 160 employees.
Elkabetz, the company’s chief executive, regularly declares that Tomorrow.io’s goal is to become the biggest weather company in the world, a boast that has annoyed market leaders AccuWeather and IBM’s The Weather Company. He testified before Congress in July, warning about some of the extreme weather events the company had spotted and asking lawmakers to invest more in weather forecasting.
This summer, the company has been buzzing cranberry bogs and sand pits around Cape Cod with a small helicopter, testing the miniaturized radar system it plans to send into space on dozens of relatively inexpensive satellites.
Ground-based radar stations provide comprehensive storm tracking in the United States, but less-developed countries such as Brazil and India have sparse coverage. Ground stations, as well as the current imaging satellites, also simply cannot provide sufficient coverage over vast stretches of the world’s oceans.
Meanwhile, NASA operates one radar-equipped satellite, called the Global Precipitation Measurement mission, that is the size of a school bus and cost $1 billion. But it provides updated views only once every few days. Some other microwave satellites help measure precipitation every few hours, but leave many gaps.
Those gaps can be critical, as the warmer climate is allowing storms to power up and intensify just hours before they make landfall. Underestimating a hurricane’s power can have devastating consequences if evacuation orders miss the mark.
Though Hurricane Henri dropped to tropical storm status before hitting the Rhode Island coast, storms making landfall in Southern states — such as Hurricane Ida — are more frequently intensifying in their final hours at sea, a change that space-based radar could catch. Indeed, New York officials said they were surprised by the intense power and heavy rains from Ida, which was blamed for causing more than 50 deaths across several states.
University of Georgia atmospheric scientist Marshall Shepherd is advising Tomorrow.io on developing its satellite network. Shepherd, who spent 12 years as a meteorologist at NASA and worked on the Global Precipitation Measurement satellite, is hopeful that Tomorrow.io can help improve both live storm tracking and longer-term models of climate change.
“If we can put a plethora of small radars up in space measuring rainfall, that will undoubtedly provide us with additional data,” Shepherd said. “A space-based network is the next frontier for improving weather, climate, and flood prediction.”
The recent warning by the United Nations’ Intergovernmental Panel on Climate Change that the atmosphere is heating up more quickly than expected highlights the need for better weather modeling to keep up with fast-changing conditions and, for example, improve warnings to people in harm’s way.
One example: A key weather system known as a mesoscale convective system, essentially large rain and lightning storms spanning 62 to 620 miles, is the cause of many extreme events, including tornadoes and flash floods. But scientists are still struggling to see what’s happening in the storms at all times and realistically include them in climate models.
Tomorrow.io’s satellites will each be about the size of a mini-fridge and cost several million dollars to build and launch into orbits of about 300 miles above the earth. The company hasn’t said much about its launch plans yet, but its options include using a private rocket company such as SpaceX or Blue Origin.
If they work as promised, the satellites could assist weather forecasters and climate modelers, said George Huffman, who oversees NASA’s weather satellite mission and runs the agency’s Mesoscale Atmospheric Processes Lab.
“Tennessee just had a 17-inch rain event; that’s a thousand-year flood,” Huffman said. “If they’re successful at helping the models, both weather models and climate models, these data could be critical.”
The company’s plan is to launch two satellites late next year and then tweak the design for the full constellation to be operational by the end of 2024.
It’s still not certain how useful the new radar data will be, said Derrick Herndon, a hurricane researcher at University of Wisconsin’s Cooperative Institute for Meteorological Satellite Studies. “There’s a theory as to what this data will provide, but the reality isn’t always the same,” he said.
“It’s a big, huge, huge challenge,” added Huffman. “They’re extremely well funded, but the technical challenges should not be underestimated.”