Data Dives are conversations with Columbia Climate School researchers to learn more about their work and explore trending topics through the lens of data science and visualization.
The Scientists
Liv Yoon‘s research is at the intersection of climate change, social inequities, and health, with a focus on community engagement work. As a social scientist and a postdoctoral research scholar at the Columbia Climate School, she approaches climate change as a sociopolitical crisis. Her research is focused on taking climate change as an opportunity to challenge the status quo and promote structural changes that alleviate social inequities that both lead to, and are exacerbated by, the climate crisis.
Joey Williams directs operations of the CAPA Heat Watch program, which provides high-resolution descriptions of urban heat based on a coordinated data-collection campaign. He combines his process design skills as a former engineer with his passion for human and environmental health to help the CAPA team advance climate resilience efforts in cities around the US and world. He earned his Master’s in Urban and Regional Planning from Portland State University, where he served as a graduate research assistant in the Sustaining Urban Places Research Lab. In his free time he enjoys backpacking, running, cooking new recipes, and drawing.
The Project
The New York City heat mapping project is part of a nationwide initiative. Williams’ team produced the heat map data shown in the map below using data collected by citizen scientists; Yoon is the principal investigator in the New York City initiative. In the following interview, they discuss how the heat map was made and what it can tell us about historical racism and present-day social inequities.
The Data
Heat Map, Northern Manhattan and South Bronx, Afternoon, 7-24-2021
Afternoon, July 24, 2021. Blue tones indicate cooler than average temperatures, reds hues warmer. Heat Data: CAPA Strategies [view data and final report]. Basemap: Esri, HERE, Garmin, Intermap, increment P Corp., GEBCO, USGS, FAO, NPS, NRCAN, GeoBase, IGN, Kadaster NL, Ordnance Survey, Esri Japan, METI, Esri China (Hong Kong), (c) OpenStreetMap contributors, and the GIS User Community
The Dive
How was this research project structured, and what data was collected? How was this raw data prepared for further analysis?
Joey Williams: Community scientists collect readings of air temperature and relative humidity using simple-to-use sensors [pdf] that attach quickly onto passenger vehicles and bicycles. The sensors automatically take measurements every one second, amounting to thousands of data points collected by community members over the course of an hour-long traverse. The sensors also record latitude, longitude, time and speed, allowing CAPA’s data scientists to pinpoint the location of each measurement along with important meta-data that help to filter out outliers and other anomalies from the raw dataset. Field notes collected by our community scientists help to validate the data and identify any interruptions to the experiment, like inclement weather.
How do you leverage high-resolution satellite imagery to describe the surrounding land use and land cover of each data point? How is this data used to inform the final heat map?
Joey Williams: In order to model ambient heat across an entire study area, CAPA’s method draws on the relationship between air temperature and the thermal properties of the urban environment. For instance, materials like asphalt and concrete tend to absorb and retain heat, whereas shaded areas mitigate air temperatures. Satellite imagery collected from the Sentinel-2 satellite constellation helps to describe the existing land and its material properties across multiple spectral wavelengths, or bands, in 10 x 10 meter square cells. This set of land variables are then combined with the thousands of ambient measurements collected by community scientists to inform a machine-learning model that makes predictions of heat across the full study area.
How does CAPA’s approach to urban heat mapping differ from other techniques?
Joey Williams: CAPA’s approach to mapping urban heat is unique in several ways. First, this approach provides descriptions of the heat we experience at the human level, or roughly one-to-two meters above the ground. The more typical approach to mapping urban heat uses satellites to directly measure the temperature of surfaces, known as land-surface temperature or LST. Second, the high resolution of CAPA’s maps allow for a more fine-grained approach to understanding the distribution of heat block to block, whereas typical LST methods provide more coarse descriptions such as at the neighborhood scale. Lastly, CAPA’s heat mapping program, Heat Watch, engages local community members and stakeholders in a process of co-creation, adding a helpful degree of ‘civic legitimacy’ to the data and models. Through the campaign process, community members are able to better understand the climate risk facing their area and what they can do about it.
As a social scientist, what comes to your mind when you look at the final heat map? What stories, past or present, does it tell?
Liv Yoon: My first reaction upon seeing the resulting maps was that I wasn’t surprised. The data confirmed previous research and satellite imagery that showed northern Manhattan and the South Bronx are hotter than other more affluent areas in our geographic area of study. But I also saw legacies of the past in the heat distribution and related disparities. For example, the image below shows our heat map overlaid on a historical map of redlining in northern Manhattan. The 1938 redlining map is a ‘residential security map’ from the Home Owners’ Loan Corporation where areas deemed ‘high-risk’ are shown in red. These were essentially Black neighborhoods, which were denied mortgage credit and development funds. To this day, these formerly redlined areas remain marginalized along race and class lines, and underserved in resources including green spaces.
Composite: Redlining and Heat Maps, Northern Manhattan
Use slider to adjust heat map opacity. Heat map has been altered to overlap with historic redlining map — areas are approximate. Redlining Map: Mapping Inequality
What was surprising was that parts of the Upper East Side were just as hot if not hotter than less affluent areas of northern Manhattan. But it’s important to remember that populations in higher income areas are more likely to have better buffers against extreme heat, such as air conditioning, working indoors in climate-controlled offices, and access to consistent, and good quality health care.
How might the heat map data be extended, or combined with other information, to produce new insights? What questions remain to be asked?
Liv Yoon: The goal of our project was to go beyond simply showing which parts of the city were hotter by presenting the heat data with other maps that depict social inequities. We want to connect the dots between extreme heat, social inequities, and health — all in the broader context of climate change as a threat multiplier. Consider, for example, the overlap with the map below that shows median household income across the city.
It’s no coincidence that the hotter parts of the city also happen to have lower incomes. Once again, this indicates that residents of these areas are less likely to have buffers against extreme heat and associated health risks, such as adequate air conditioning and good quality health care.
These areas also show higher rates of chronic diseases such as hypertension and diabetes, which increase one’s chance of developing a heat-related illness.
Map: 5-Year Average Annual Heat Stress Hospitalizations, Age-Adjusted Rate
Data: NYC Environment & Health Data Portal
It’s important to remember that residents of these areas are not ‘biologically wired’ to be more susceptible to these diseases. Rather, as with so many health conditions, there are social and environmental determinants at play here. For example, lower quality food and health care, and the weathering effect (the idea that stress manifests physiologically, with disadvantaged populations experiencing this effect more acutely) are some of the social determinants that could lead to such stark overlaps.
The impact of extreme heat is also exacerbated by poor air quality — another environmental problem often found in low-income communities. Poor air quality is a chronic problem in the South Bronx, and particularly in the Mott Haven-Port Morris area, which is known as ‘Asthma Alley‘ due to having one of the highest childhood asthma rates in the country.
Map: Asthma Emergency Department Visits, Children 5-17 Years Old
Estimated Annual Rate (per 10,000 residents), 2018. Source: NYC Environment & Health Data Portal
Polluting industries, high traffic, and a lack of green space all contribute to poor air quality in this area. High heat, smog and poor air quality all compound the risk of asthma flare-ups.
We can also see that these hotter, low-income areas also have less green space.
Map: Tree Canopy Cover Percent
Data: NYC Environment & Health Data Portal
The COVID-19 pandemic shed light on the compounding and cascading problems associated with a lack of green space. Early in the pandemic, the public was advised to seek refuge from the virus outside and to maintain good physical and mental health. But healthy green spaces were not an option for many in the Mott Haven-Port Morris area, where the outdoors is merely industry, waste transfer facilities, and highways. Such advice is even less helpful when the effects of extreme heat are added to the list.
One question that remains is the nature of extreme indoor heat. For instance, residents of these areas told us that their homes retain heat from the day, so they don’t get relief even at night when the air outdoors cools down. This can cause poor sleep and fatigue which impacts one’s overall health and well-being. While indoor heat was beyond the scope of our project, organizations like WEACT have been working with community members on this issue.
How can this data be leveraged to produce meaningful change in communities? What are the next steps?
Liv Yoon: Below are some project initiatives that came about from consultation with South Bronx Unite:
- We are producing an online story map that connects the dots between extreme heat, social inequities, and health — all in the context of increasing urban populations and climate change as a threat multiplier. We also plan to include maps of various social indices such as demographic data, health disparities data, and other relevant socio-environmental determinants to explore the overlaps and connections. Our aim is to have this complete before the heatwaves this summer.
- Christian Braneon, co-investigator on this project, and I are both members of the New York City Panel on Climate Change (Christian is co-chair.) The group is working on its fourth assessment so this recent granular data will be configured into the report.
- In collaboration with NYC Department of Health and Mental Hygiene, we hope that our data can contribute to bolstering the city’s efforts at combating heat-related disparities. For example, our street-by-street granular data can help the city locate additional cooling centers or green spaces in most-affected areas, and programs such as allocation of free air conditioning for disproportionately affected communities.
- South Bronx Unite has been advocating for more green and blue space in their neighborhood. They plan to use this recent data to help push their efforts.
More about the project: Study Maps Urban Heat Islands With Focus on Environmental Justice
Interactive: Heat Story NYC
Data: Heat Watch Bronx & Manhattan (2021)
More on CAPA’s heat mapping methodology: Integrating Satellite and Ground Measurements for Predicting Locations of Extreme Urban Heat
Be sure to leave your questions and ideas for future Data Dives in the comments section below! ↓
hard to know what this kind of presentation means, since cannot see actual thermal signatures (specific temperature, time(s) of day, &c), – what actual temperature (or ranges) may be, so cannot be readily used for further analysis. A bunch of interpretations, but where are the fundamental parameters/variables mapped?
If you click on view data in the captions of most of the maps, you’ll find all of those things.