Mitigating the urban heat island effect in Missoula, Montana
Which neighborhoods in Missoula could most benefit from interventions to reduce heat?
Imagine a sunny, hot day in mid-August. Now think about standing in the middle of the Best Buy parking lot out on Reserve Street, with heat radiating off the endless pavement and buildings.
Now use the swiper to magically transport yourself under a bunch of shady trees in the middle of the University District. Big difference, right?
On the hottest days in cities, people living in areas with few trees, little green space, or lots of paved surfaces can experience up to a 20 degree difference in temperature from communities in shadier areas or places with less pavement.
These hotter zones are known as Urban Heat Islands (UHI) and are a byproduct of our built environment. Cars, air conditioning units, concrete parking lots, and buildings both generate and radiate heat.
In all, these aspects combine to create islands within our built environment where temperatures are hotter than their surroundings.
And this has a big effect. Extreme heat is the deadliest climate impact, causing at least 1,100 deaths each year in the U.S. — more than any other weather-related hazard. Heat kills directly, through heat stroke; and indirectly, by making chronic conditions such as cardiovascular disease or diabetes worse.
Missoula is relatively cooler than much of the United States, yet UHIs here can still negatively impact physical and mental health, lead to more expensive cooling costs for homes and buildings, and contribute to poor air quality through increased energy use for cooling.
This is especially true for vulnerable groups like the very old or very young, unhoused individuals, and people living in poverty.
One of many ways that communities can take action to reduce the urban heat island effect is by planting more trees. On average, urban forests are about 3 degrees (F) cooler than unforested urban areas.
To decide where tree planting or other strategies are most appropriate, officials first needs to know three things:
1) where the worst effects of extreme heat are being felt,
2) where most people live, and
3) where tree cover already exists.
Let's dive in.
Setup & Approach
The goal is to create an urban heat island index for Missoula showing where to prioritize tree planting or other adaptation or mitigation strategies. I used three variables to create the index: high average summer surface temperature, percent of area without tree cover, and population density.
Variables
Average Summer Surface Temperature
The first variable in the index is high average summer surface temperature. By comparing this data against the Missoula area, I can calculate the highest average summer surface temperature for each Census block — the smallest spatial unit that the Census collects data for.
Lighter colors correspond to lower average summer temperatures; darker purples correspond to the higher temperatues.
Zooming to the core of Missoula, we see an eye-catching demarcation separating the western and eastern areas of the city. Generally, areas west of Bancroft Street show noticeably hotter average temperatures.
Tree Cover
Missoula's city limits (in white) mark the boundary of the project's focus. Tree cover is displayed in green.
Even at this scale, differences in tree cover are visible across the city. Can you pick out the sparse tree canopy over the commercial and industrial areas on the north side? How about the dense trees lining the Clark Fork river as it cuts east to west through the heart of the city? Or the abundant tree canopy shading the University District in the southeastern corner?
Here, darker green indicates denser tree cover and the lighter green and white indicate less cover. For instance, the blocks displayed in white are those with 89% or more of their area not under tree cover, while the darkest green blocks have less than 30% of their area uncovered.
Population Density
The last variable for the index is the population density for each Census block in Missoula. This component of the index prioritizes populated areas where the most people may benefit from interventions, and using population density rather than raw population counts ensures parity across Census blocks.
Darker purples correspond to denser populations.
Creating a Heat Risk Index
There are several ways to calculate a composite index, and each approach has pros and cons that make it more or less appropriate given the overall context and analysis goal.
Compensability is one of those nuances; a complex-sounding word that simply refers to variables in the index with high values compensating for other variables with lower values, creating higher overall index values that may not align with the intent of the analysis.
To avoid, for instance, a block with a dense population, a pronounced lack of tree cover, but relatively low average summer temperatures from receiving a high index value, I chose a multiplicative approach — using the geometric mean to combine the three input variables — that negates this kind of compensability and ensures higher index values in areas where all the input values are high.
The geometric mean is appropriate in this situation as it deals well with skewed data (e.g. the population density and tree cover variables) and is indifferent to variables on different scales.
Categories
Heat Risk Index (HRI) values are grouped into five different categories, organized from the areas projected to least benefit from heat interventions to the areas most in need.
The Least Benefit category consists of many outlying, unoccupied, and sparsely-populated areas around Missoula.
The Low Benefit category represents roughly 15k people, and these blocks have on average 57% of their area covered by trees — though this average is inflated somewhat by a few blocks that mostly consist of forested land with only small portions of their area inhabited.
The Moderate Benefit category holds the largest population, with nearly 37k people.
The High Benefit category holds more than 28k people and is made up of much of Missoula's industrial areas.
Finally, the Highest Benefit category includes seven Census blocks that represent some 800 people who live in the hottest areas with the least tree cover.
These seven blocks constitute a mixture of residential and combination residential / industrial areas throughout the city.
Ideally, this work can catalyze a conversation between the City of Missoula, local social services, & conservation organizations about how warming impacts Missoulians and what interventions might be cost-effective and appropriate for different areas.
Officials could use this map, together with additional research and mapping efforts, to prioritize a great number of interventions, such as...
Urban tree-planting and pocket parks that not only reduce ambient temperature, but can reduce physical and mental stress, boost moods, and decrease the severity of illnesses like anxiety and depression .
Building shaded structures on trails heavily exposed to the sun.
Incentivizing cooling strategies such as cool roofs , green roofs like these seen in the background image, energy audits and efficiency projects, and other green building infrastructure.
Problems from climate change build on one another, like when extreme heat worsens air quality. But the solutions can have positive synergistic effects, too.
Cooling the city with shade trees and light roofs, for example, will result in lower energy use and reduced greenhouse gas emissions — a win for reducing emissions, for healthy people, and for strong communities.
Interactive Map
Now that the stage has been set, the gloves are off for you to explore the interactive map yourself. What patterns do you find?
Heat Risk Index (HRI) values are grouped into five different categories, organized from the areas projected to least benefit (lighter red / white) from heat interventions to the areas most in need (dark red). You can click the legend in the upper right hand corner to view the categories, and also toggle the city blocks layer and city boundary layer on or off.
Future Work
I plan to improve this passion project with additional analytical rigor in the coming weeks, as time allows. Some areas that would be most beneficial to integrate are:
- Adding a JEDI (justice, equity, diversity, inclusion) lens to the analysis by integrating a measure of vulnerability to heat. Several studies (for instance, here ) have found that some communities in the United States, particularly those that are low-income and with higher populations of people of color, have neighborhoods with higher temperatures relative to adjacent neighborhoods in the same city. The studies identify historic redlining — a now-illegal practice from the 1930s when the federal government labeled non-White neighborhoods as undesirable for real estate investment — as a contributing factor. Specifically, people of color and community members with low incomes are more likely than other groups to live in historically redlined neighborhoods that are present-day intra-urban heat islands. I can also see age, disability status, availability of care or other interventions, and other illnesses as potential additional factors contributing to heat vulnerability — e.g., the EPA's work in this field, here.
- Extrapolating Block Group-level data to the smaller Census Blocks. Several of the indicators above, for instance income, are not available at the granularity of Census Blocks. I'll need to research an rigorous way to apply BG-level socioeconomic data to the individual Blocks if I'd like to continue working at Block-level granularity. Given the relatively small scale of Missoula, I believe Block-level detail is best for this analysis.
- Using a more nuanced temperature variable. Currently my temperature variable is based on the highest average summer (June, July & August) air temperature reading. This may not fully take into account temperature differences between blocks due to impervious and heat-reflective surfaces such as pavement, glass, and metal roofs. A more nuanced temperature variable would be to calculate the highest land surface temperature for each block, rather than air temperature.
- Adding validation to the analysis, such as hospital data on heat-related admissions or community survey data on perceived heat vulnerability. Ultimately, the goal of this analysis is to create a decision-support tool to help city planners equitably and effectively address heat vulnerability across Missoula. That's a serious mission, with real lives at stake. That demands rigor and outside accountability to ensure that the tool is truly fulfilling its mission.
