AI is powering Google’s new heat resilience tool

As summers get hotter, the insidiousness of extreme heat becomes more evident — even as many continue to underestimate its danger. For investors and other major players in climate resiliency, this represents an opportunity for innovation.

Google is the latest, today unveiling a new heat resilience tool, which applies artificial intelligence to satellite and aerial imagery to provide cities with data about how hot they’re getting. 

Mansi Kansal, a research product manager at Google who worked on the tool, told Latitude Media that the hope is that cities will leverage the data to implement cooling interventions such as cool roofs, which have been found to reduce peak energy demand for cooling by up to 27%. 

The heat resilience tool gives access to a detailed map with information about average summer surface temperatures, census tracts, and the presence or absence of trees and cool roofs in any given neighborhood. Using it, a city could target its cooling interventions toward the places that need them the most. 

“One of the barriers cities face is limited funding,” Kansal said. “We hope that using our quantifiable data, cities can [see] how to reduce temperatures by two degrees, three degrees, one degree, even a 10th of a degree, and [use the data to] apply for grant funding.” 

Google is currently piloting the tool in 14 cities, including Miami-Dade County, which plans to use the information to develop policies to encourage developers to take heat mitigation measures. 

“[The tool] was a combination of us realizing that Google has a lot of really unique data through our maps [and] the imagery that Google Maps provides, and this need for cities to understand how they can implement strategies for addressing heat,” Kansal added. 

A social responsibility to adapt

The launch is part of a renaissance of climate resilience tools that has seen startups providing everything from wildfire monitoring to climate risk assessments pop up across the climate tech space. 

It comes in part as a result of investors in advanced economies finally coming to terms with the fact that we’ll need to adapt to the consequences of climate change — regardless of our investments in mitigation. But the rapid advancement is playing a part as well. 

“In the past, you could do [understand the land characteristics] at a lower resolution and just at an area level,” Kansal said. “But with our AI improvements, we can understand the roof reflectance of individual roofs. Combining that with regression models allows us to do this type of modeling.”

The heat resilience tool is the latest addition to Google’s climate adaptation assets, which include flood monitoring and wildfire tracking. The idea for the suite of adaptation tools, the story goes, came over a decade ago, when a Google employee saw a wildfire from the window of his office in Israel and realized there was no easily accessible information on how to monitor it. 

Unlike floods and fires, however, heat is a more subtle kind of climate disaster, a “silent killer,” as Kansal puts it. 

“We see hurricanes, we see floods, they are very visible,” she said. “Whereas heat is pervasive, and people don't realize how much it impacts human health.”