Amid all of the discourse on the devastating consequences of climate change, one fact often slips through the cracks: Targeted interventions can reduce potentially deadly heat exposure for city residents, not just in years to come but for people living in the present day.
Examples can be as simple as painting roofs white to reflect the sun’s rays or adding bus shelters to shade commuters. Others require ongoing attention, such as making sure newly planted trees stay healthy. Since every city and community is different, there’s no one-size-fits-all set of solutions, experts told the Tribune. But there is a universal starting point: Identifying areas where temperatures are often hottest.
Knowing whether a community tends to have elevated temperatures makes a difference, especially for people more sensitive to heat stress, including children, older adults, athletes and people who are low-income, are pregnant, have certain medical conditions or work outdoors. Other cities have made this information accessible. But Chicago has not.
In setting out to identify Chicago’s hottest areas, the Tribune realized that data from regular temperature sensors on the ground — such as those at O’Hare and Midway airports — couldn’t paint a complete picture. Temperatures can vary from community to community — sometimes from block to block.
Turning to satellite data
Tribune reporter Sarah Macaraeg first became familiar with remote sensing data, or observations of Earth gathered by satellites, in previous coverage of a community with insufficient air quality monitoring stations. Extracting temperature data and other information from U.S. Geological Survey satellite images proved difficult, however, until the Tribune turned to Google Earth Engine.
A Google portal not widely known outside of climate research circles, Google Earth Engine allows users to filter data about the environment by specific measures and locations, using the JavaScript or Python programming language, as well as to run calculations via Google servers with far more processing power than the ordinary computer.
After referencing Google Earth Engine’s code library and scientific studies, and checking best practices for using filters to obtain the highest-quality data, the Tribune used Google Earth Engine to gather land surface temperature data from the Landsat-8 and Landsat-9 satellites, which orbit around 500 miles above the Earth’s surface, passing over the same location every 16 days.
Heat in Chicago: Explore how surface temperatures vary
Unlike daily readings from the MODIS satellite, which are also often used in heat mapping, Landsat satellites gather observations in smaller increments — 30-meter by 30-meter areas — that are better suited to illustrating community-level differences. The Tribune used Google Earth Engine to match readings taken from these areas to census block groups and to translate the readings from degrees Celsius to degrees Fahrenheit.
Because clouds affect the accuracy of remote sensing data, the analysis was based only on observations gathered on summer days with minimal clouds, from the first year Landsat-8 data was available, in 2013, to the end of August 2022, when Landsat-9 also was collecting data. Per standard scientific practice, specific data points USGS had cataloged as “cloud” or “cloud shadow” were omitted, along with those marked “water.”
The Tribune’s analysis was based on average temperature deciles, which are measurements in increments of 10 that show how values relate to one another — for example, the 10th decile representing block groups cooler than 90% of the rest of Chicago — along with race, ethnicity, poverty, household and health care data from the U.S. census. The data tables were joined and analyzed using Structured Query Language.
The Tribune then created an initial map in ArcGIS and used it to begin outreach to residents of historically hotter neighborhoods and to inform the location of on-the-ground reporting during heat advisories last summer. Former Tribune reporter Stephanie Casanova and Tribune photographer E. Jason Wambsgans teamed with Macaraeg for this reporting.
Because exposure to extreme heat can exacerbate the health impacts of air pollution and vice versa, the Tribune also analyzed the Chicago Department of Public Health’s Air Quality and Health Index, which provides percentile measurements of air quality as of 2020 in about 2,170 block groups with sufficient data.
To better understand the allocation of some cooling resources, the Tribune used data from the Chicago Transit Authority and the city to identify the locations of bus stops that were in service in the city as of March 2022 and the location of bus shelters in Chicago as of October 2021. The Tribune requested updated bus stop and bus shelter location data in April under the state Freedom of Information Act, but the CTA did not provide it. The location and acreage of public parks in Chicago were also merged in ArcGIS, joining the location of the center of a park to its corresponding block group to estimate their distribution by decile.
Partnering with researchers
To further refine, verify and visualize the data analysis, the Tribune partnered with Boston University’s Center for Climate and Health, whose researchers also added an analysis of access to cooling resources in Chicago’s most vulnerable communities with the hottest average surface temperatures.
Muskaan Khemani, a researcher who’d mapped heat in Wichita, Kansas, as part of an internship with NASA, analyzed the data anew in Google Earth Engine, removing observations gathered within a kilometer of water in addition to those cataloged as clouds, cloud shadows or water. Analyzing average summer surface temperatures by block group in the programming language R, Khemani found the same disparities as the Tribune did.
The final analysis of surface temperatures is based on 35 images from the Landsat-8 and Landsat-9 satellites gathered on days with less than 20% cloud cover from June 1 through Aug. 31 in the years 2013 through 2022, and five-year census estimates from the 2021 American Community Survey.
Of 2,313 census block groups with population data available that fall fully or partially within city of Chicago boundaries, 10 didn’t have sufficient temperature data to be included. Boston University’s analysis of racial and ethnic disparities incorporates the population of some block groups whose boundaries extend beyond the Chicago border. Estimates related to demographics and resources in the city’s hottest and coolest areas do not.
Of course, temperatures don’t strictly align to block group boundaries. But it’s important to map heat differences within recognizable boundaries to help allocate resources, Khemani said. “It allows you to see some variability in surface temperatures, and that spatial unit corresponds to something real and tangible for decision-makers.”
Experts who spoke with the Tribune noted a crucial caveat about land surface temperatures: They are not the same as the air temperature measurements seen in weather reports. As Chicagoans familiar with summer’s scorching sidewalks intuitively know, land surface temperatures are more extreme — hotter when it’s hot and cooler when it’s cool. “It won’t be quite as exact or precise like an on-the-ground station, just because it’s measuring something that’s not exactly (air) temperature. But it does give you a spatial pattern,” Illinois State Climatologist Trent Ford, who did not participate in the analyses, said of remote sensing.
Khemani calls Google Earth Engine a “game changer” that shouldn’t be seen only as the province of full-time researchers. “This was not feasible a decade or two ago because there was just no way to download as many images as we’re using now. It’s such a valuable tool for researchers and people outside of researchers to use,” she said. “So many more people and organizations are able to conduct these types of analyses.”
For those not interested in learning the coding skills needed to make use of Google Earth Engine, Boston University researcher Jason Rundle said there’s still a wealth of other population, health and environmental data that reveal insights without heavy-duty analysis. The Centers for Disease Control and Prevention’s Social Vulnerability Index, for example, reflects a combination of measurements including race, poverty status and disability at the census tract level.
“It’s important to not just keep written scientific research in this box where only a certain type of person can do it or you have to have this many degrees,” Rundle said. “Even just looking at data in your own neighborhood is useful and you can learn so many things that you didn’t know before.” Chicagoans interested in further researching their census tract or block group can access those identifications by searching the interactive map published by the Tribune and clicking through the details that pop up.
Rundle analyzed surface temperatures in conjunction with the CDC vulnerability index and the location of all public cooling resources promoted by the city of Chicago — from its six official cooling centers to the Park District, police station and library locations it also advertises during heat advisories. Using ArcGIS’ network analysis tool, Rundle identified numerous areas that have the highest social vulnerability rating, the hottest surface temperatures and no public cooling within walking distance of a half-mile.
He and Khemani both said they hope to see other researchers and journalists pursue similar partnerships in the future, to cover local aspects of the climate crisis in depth. Partnering with journalists helps make climate science more accessible to the public, they said; partnering with scientists can bolster the strength of journalists’ analytical work and provide institutional resources that newsrooms may lack to host and visualize large data sets.
With multiple experts stressing the importance of community member preferences in determining interventions, the Tribune attended online town hall meetings on the city’s climate action plan, reviewed city data on green roof permits and researched community initiatives to identify residents already attempting solutions in their neighborhoods.
“It doesn’t always have to be these doomsday scenarios,” Rundle said of messaging about the climate crisis. “There’s hope there, and there’s people learning how to thrive as these things are happening.”
smacaraeg@chicagotribune.com
