Current inequality and future potential of US urban tree canopy cover for reducing heat-related mortality, morbidity and electricity consumption

doi:10.21203/rs.3.rs-2868700/v1

Excessive heat is a major and growing risk for urban residents. Urban trees can significantly reduce summer peak temperatures, thus reducing heat-related mortality, morbidity, as well as cooling energy demand. However, urban tree canopy is inequitably distributed in US cities, which has been shown to contribute to higher summer temperatures in people-of-color (POC) neighborhoods. Here, we utilize a unique dataset of high-resolution tree canopy cover to estimate the inequality in summertime heat-related mortality, morbidity, and electricity consumption across a sample of 5,723 US municipalities and other places, housing 180 million people during the 2020 census (50.6% in majority non-Hispanic white neighborhoods, 49.4% in majority people-of-color, POC, neighborhoods). We also model the potential to increase tree cover in these municipalities in 5% increments up to a realistic maximum, the 90th percentile of tree cover observed in each municipality for each impervious surface class. On average, trees in non-Hispanic white neighborhoods cool the air by 0.19 ± 0.05⁰C more than in POC neighborhoods, leading annually to trees in white neighborhoods helping prevent 190 ± 139 more deaths, 30,131 ± 10,406 more doctors’ visits, and 1.4 ± 0.5 terawatt-hours (TWhr) more electricity consumption than in POC neighborhoods. The greatest inequality in the protective value of trees occurs in the Northeastern US. We estimate that for these 5,723 municipalities, the maximal feasible urban reforestation program in residential neighborhoods could add 1.2 billion trees, reducing population-weighted average summer temperatures by an additional 0.38 ± 0.01⁰C. Relative to the current status quo, this increase in the cooling benefits of trees would reduce annual heat-related mortality by an additional 464 ± 89 people, annual heat-related morbidity by 80,785 ± 6110 cases, and annual electricity consumption by 4.3 ± 0.2 TWhr, while increasing annual carbon sequestration in trees by 23.7 ± 1.2 MtCO₂e and decreasing annual electricity-related GHG emissions by 2.1 ± 0.2 MtCO₂e. The total economic value of these benefits, including the value of carbon sequestration and avoided emissions, would be USD 9.6 ± 0.5 billion, although in many neighborhoods the cost of increased tree cover exceeds these benefits. The exception is neighborhoods that currently have lower tree cover, especially those that are majority POC, as these neighborhoods have a much higher return-on-investment from tree planting.

Earth and environmental sciences/Ecology/Urban ecology

Earth and environmental sciences/Climate sciences/Climate change/Climate-change impacts/Environmental health

Biological sciences/Ecology/Urban ecology

Earth and environmental sciences/Environmental social sciences/Climate-change adaptation

Scientific community and society/Forestry

Climate change is already here, with myriad impacts on human health and society that are projected to increase over time as changes intensify^1,2. One important impact of climate change is increased risk from excessive heat, with increases in average summer temperature as well as the frequency and intensity of heat waves. Already, in an average year, heat stress kills roughly 6,100 Americans and 356,000 people globally³, as measured by epidemiological studies of excess deaths due to high temperatures. Higher air temperatures increase mortality and morbidity by causing heat stroke and exhaustion, and by exacerbating existing cardiovascular, pulmonary, and renal diseases^4,5. Summer temperatures in the United States (US), the focus of this paper, are increasing and heat waves are becoming more frequent as climate change intensifies, with the average number of days with a dangerous heat index (high temperature and humidity) forecast to triple by 2050⁶.

Projected impacts on mortality are uncertain, varying among studies. One study projected that by 2090-2099, climate change in warmer regions of the world could increase annual heat-related mortality by 3.0% to 12.7%, depending on the region and the greenhouse gas (GHG) emissions scenario used, with annual heat-related mortality in the US forecast to increase by 3.5% under the RCP8.5 emissions scenario⁷. By contrast, another study in just the Eastern US projected an additional 11,562 annual deaths by 2050, which would represent a large increase in annual heat-related mortality⁸. While these two examples from the literature illustrate the wide range in projections of health impacts from increased heat waves, there is general agreement in the literature that climate change in the US will make heat waves more frequent and intense, increasing mortality and morbidity^9,10.

Trees can reduce solar radiation hitting surfaces such as asphalt and concrete, reducing temporary heat storage in these materials, which would later be emitted as thermal radiation. Trees also lose water to transpiration as they respire, and this phase change of water from liquid to gas increases latent heat flux¹¹. The net effect of trees is therefore to reduce ambient air temperature, particularly during daytime. The effect of trees is fairly local, reducing air temperatures primarily within a few hundred meters horizontal distance from the tree canopy, although this distance varies with factors such as the size of the canopy patch¹², wind speed, and relative humidity⁵. It is important to note that while air temperature is correlated with health impacts, there are other heat stress metrics that take into account factors also important for thermal comfort like humidity¹³ and direct solar radiation^14,15.

Heat action planning has been used by governments at many scales to reduce the risk to vulnerable populations. Common components of heat action planning include the creation of early warning systems, cooling centers, and response plans by medical institutions^16,17. One part of heat action planning can be actions to reduce ambient outdoor air temperature^18,19, and many city governments have begun to consider the additional role trees can play to reduce ambient outdoor air temperatures^20-26. Moreover, trees already provide significant heat-reduction benefits⁵. One study in the United States estimated that without urban tree canopy, annual mortality would have been 1,200 deaths greater than at present²⁷.

Research on urban tree cover shows its distribution is quite unequal. One way to measure this is to compare tree canopy across US Census blocks, for which we have associated socioeconomic data. In most US cities, low-income or people-of-color (POC) neighborhoods have less tree cover than high-income or non-Hispanic white (hereafter simply “white”) neighborhoods^28-37. One national survey of 5,723 municipalities in the US (the same set of municipalities used in this study) found that in 92% of communities, low-income blocks have less tree cover than high-income blocks, with low-income blocks on average having 15.2% less tree cover and being 1.5⁰C hotter (summer surface temperature) than high-income blocks³⁸. Multiple other studies show similar results using a variety of different methodologies, consistently finding that in most US cities, lower income and minority populations live in neighborhoods with higher summer surface temperatures^39,40.

Beyond the health impacts, heat has other impacts on society. For instance, Santamouris and others⁴¹ found an average increase of 0.45-4.6% in peak electricity load for each 1^°C increase in air temperature. Trees near buildings can be one way to decrease electricity use, among other strategies such as implementing energy efficiency programs, reducing building albedo (“cool roofs”), and creating incentives to conserve electricity⁴². Tree cover reduces ambient air temperature and shades buildings from solar insolation, reducing the energy demand for indoor cooling from 2.3-90% depending on the study⁴³. This avoided electricity consumption also avoids greenhouse gas emission from electricity generation.

Urban trees also sequester and store carbon⁴⁴, among other ecosystem services⁴⁵. Urban trees are one kind of natural climate solution⁴⁶, defined as conservation, restoration, or management actions that increase carbon sequestration or avoid greenhouse gas emissions. Planting new trees in urban areas (afforestation or reforestation depending on site history, but for simplicity referred to as reforestation in this paper) will lead to net carbon storage as the trees grow, even after accounting for GHG emissions associated with tree planting and maintenance^47,48.

Previous estimates of the urban reforestation potential in the United States were designed to estimate the maximum carbon sequestration potential. For instance, Fargione et al.⁴⁴ assumed that essentially all land in urban areas not currently developed (e.g., impervious surfaces, buildings) or under other land-use (e.g., agriculture, sports fields or golf courses) could be available for tree planting, estimating that 23.3 (19.17-30.15 95% CI) Mt CO₂/yr of carbon sequestration was possible, around 1.9% of the total NCS potential in the US⁴⁴. Similarly, Cook-Patton et al.⁴⁹ modeled fully planting all low-density open space in urban areas across the contiguous US where forest naturally occurs, and estimated that 52.5 Mt CO₂/yr of carbon sequestration was possible. A recent study using high-resolution imagery of the US state of California estimated that 4.5 Mt CO₂/yr from urban reforestation was possible in that state⁵⁰.

However, only a fraction of potentially available urban open area is likely to be plantable, due to municipal codes and land-use regulations, landowner preferences, and conflicts with other land-uses that require sparse vegetation like lawns. Moreover, past efforts at estimating the NCS potential of urban reforestation often have been based off 30m Landsat-derived tree cover data. These coarser datasets miss many single tree canopies in urban areas and can significantly underestimate current tree cover⁵¹, particularly in landscapes with less than 30% tree cover⁵². Underestimation of current urban forest canopy may lead to overestimation of reforestation potential.

While there have been studies of individual cities or sets of cities^20-25, there has been no quantitative national prospective study in the US of a large sample of thousands of cities to estimate the potential that increases in urban tree cover have to reduce mortality, morbidity, and electricity consumption. To our knowledge, there is also no national assessment of how the potential of trees for heat-risk reduction varies across gradients of race/ethnicity. Understanding the potential of urban trees to serve as an adaptation to more frequent and intense heat waves is particularly important at this moment, when there is substantial investment in NCS as countries plan to achieve their Nationally Determined Contribution (NDC) commitments under the Paris Climate Accords⁵³, and the US has committed to increased funding for climate-related forestry programs as part of the Inflation Reduction Act⁵⁴. Similarly, substantial investment in climate adaptation is likely in the next decade by entities like the Green Climate Fund⁵⁵. Well-targeted, cost-effective investment in nature-based solutions for climate mitigation and adaptation is difficult without spatially-detailed data on reforestation potential and need⁵⁶.

In this paper, we use a high-resolution (2m) tree cover map for all 100 US urbanized areas > 500 km² which contain 5,723 municipalities or other Census-designated places, and housed 180 million people in 2020 (68% of the people who live in urbanized areas in the United States or 55% of the national population) ³⁸. For this large sample of 5,723 municipalities (50.6%, or 91 million population, in majority white neighborhoods, 49.4%, or 89 million population, in majority POC neighborhoods), we estimate reforestation potential at the Census block-level, accounting for variation in impervious cover among blocks and assuming that reforestation is only likely up to levels commonly seen in other blocks of similar impervious surface cover. We estimate the change in air temperature due to tree planting and the resultant reductions in human mortality, morbidity, and electricity consumption. In addition, we estimate the carbon mitigation potential achievable through this large-scale urban reforestation scenario. Across our analyses, we propagate statistical uncertainty to describe the precision of our estimates. The major goals of this paper are to:

Quantify the inequality in the protective value trees currently provide by reducing mortality and morbidity, across income and race/ethnicity.
Estimate, for a variety of planting scenarios, the increase in urban tree canopy, the number of trees planted, the net carbon sequestration, and the costs. For each urban reforestation scenario, quantify the likely reduction in summer air temperatures, as well as the associated reduction in mortality, morbidity, and electricity consumption.
Examine the return-on-investment (ROI) curve of urban reforestation’s effect on avoided mortality, morbidity, and electricity consumption, to test whether ROI varies systematically with neighborhood income or race/ethnicity.

Current national benefits:

On average across the country, neighborhoods with a majority of people of color (POC) have 11% less tree canopy cover and 14% more impervious surface than white neighborhoods (Table 1; unless otherwise stated we present in this paper the difference expressed as percentage points, %p). We estimate that the current tree canopy on average reduces air temperatures by a population-weighted 1.01 ± 0.03⁰C in white neighborhoods compared to 0.82 ± 0.03⁰C in POC neighborhoods.

Table 1

**Inequality in tree canopy cover, air temperature, and heat-related mortality, morbidity, and electricity use.** The median reduction in air temperature and total reduction in impacts is shown relative to a hypothetical case with no urban tree canopy cover, to quantify the ecosystem service that urban tree canopy cover provides to heat-health. Neighborhoods are divided into those that are predominately non-Hispanic white, or those that are predominately people of color (POC). Confidence intervals shown are calculated from regression analysis and error propagation, see Methods for details. Tree cover and impervious surface cover are measured using remote sensing, and so the mean for our sample of cities is known with high precision (< 0.1%), see Methods for a discussion of accuracy of classified remotely sensed imagery. Totals presented are for the 5,723 municipalities in our sample.
				Reduction due to trees in annual heat-related impacts
Race/ethnicity	Tree Canopy Cover	Impervious Surface	Median reduction in summer air temperature due to trees (⁰C)	Mortality	Morbidity	Electricity (TWhr)
Majority non-Hispanic white	35%	42%	1.01 ± 0.03	632 ± 100	114,936 ± 7,444	6.2 ± 0.3
Majority POC	24%	56%	0.82 ± 0.03	442 ± 97	84,805 ± 7,271	4.8 ± 0.3

We estimate that in majority white neighborhoods trees annually help avoid 632 ± 100 deaths. The annual economic value of that avoided mortality is USD 4.5 billion (Fig. 1). In contrast, we estimate that, even though the total population in each race/ethnicity type are of relatively equal size (91 million in white neighborhoods vs 89 million in POC neighborhoods), in majority POC neighborhoods trees help avoid 442 ± 97 deaths annually, an annual economic value of avoided mortality of USD 3.1 billion. This differential in benefits extends to other types of benefits. Morbidity avoided because of tree cover is 30,131 ± 10,406 greater in white neighborhoods than in POC neighborhoods (Table 1). Similarly, shading from trees reduces total electricity consumption in white neighborhoods by 1.4 ± 0.5 TWhr more than it does total consumption in POC neighborhoods (Table 1).

Patterns within urbanized areas

There is significant variation in current tree cover within urbanized areas, which leads to variation in the potential for tree cover increases to provide benefits. To demonstrate typical trends, we provide a case study for Washington DC in Fig. 2. At a neighborhood scale, there is significant variation in tree cover, driven by differences in settlement density and urban form, as well as the existence of patches of forests within protected areas or otherwise undevelopable sites (Fig. 2, upper left). Neighborhoods in DC that predominately house people of color have higher population density (average of 4,828 people/km²) and are more often centrally located, than neighborhoods that are predominately comprised of whites and are more often suburban or exurban (3,962 people/km²). On average, tree cover is 12% lower in POC neighborhoods than in white neighborhoods due to the correlation with population density (and hence impervious surface cover) ³⁸ as well as other historical factors, such as redlining⁵⁷.

Across the entire Washington, DC, urbanized area, there is an overall gradient in summer land surface temperature (Fig. 2, upper right), with less densely populated suburbs and exurbs having lower land surface temperature than dense urban core neighborhoods, due to their lower impervious surface cover and higher tree cover. This pattern means that the greatest reduction in heat risk from additional planting, measured as annual avoided deaths per million people (M), occurs in denser, often POC, neighborhoods (Fig. 2, lower left). However, the tree planting potential, measured in number of additional stems or in potential carbon sequestration, is greatest in suburban and exurban areas (Fig. 2, lower right), which have low cover of impervious surfaces and hence more space for planting, even though those are the neighborhoods that already have disproportionately high tree cover.

Patterns among urbanized areas

There is also variation among urbanized areas in the amount of tree cover and the protective benefits trees are providing from heat-related mortality and morbidity. The largest benefit in terms of deaths avoided is in the urbanized areas with the greatest population, such as those centered around New York City and Los Angeles (Fig. 3). This is not surprising, since there is a greater population potentially at risk from heat in the largest urbanized areas, and so in absolute terms the number of avoided deaths due to trees is greatest in these urbanized areas. Other factors matter as well, however. Urbanized areas that are densely settled, with a lot of impervious surface cover, and yet have significant tree cover will have a relatively high number of avoided deaths due to trees, all else being equal. For instance, Boston (population 4.5 million) has both high impervious surface cover and relatively high population density, and trees help avoid an estimated 54 deaths annually. In comparison, Atlanta (population 5.1 million) has lower impervious surface cover and lower population density, and trees help avoid an estimated 32 deaths annually.

However, the protective benefits trees provide in terms of avoided mortality are unequally distributed with respect to race/ethnicity. To account for differences in population between majority POC and majority white neighborhoods, we calculated the protective rate, the number of heat-related deaths avoided annually due to trees per million people. The protective rate is generally lower in majority POC neighborhoods than in majority white neighborhoods (urbanized areas with yellow to red colors in Fig. 3). The inequality in protective rate is greatest in the Northeast of the US, in the so-called Northeast Corridor stretching from Washington (DC) to Boston (MA). These cities have substantial inequality in tree cover and surface temperature among neighborhoods³⁸, which leads to a substantial inequality in air temperature and the protective benefits of trees in reducing heat-related mortality. Northeastern US cities are characterized by dense city centers, often formed prior to the widespread availability of the automobile, which often contain many of the POC neighborhoods, and lower-density suburbs and exurbs that are often majority white. Conversely, there are urbanized regions where the protective rate is greater in POC than in white neighborhoods, such as Southeastern US cities like Atlanta. In these cities, most neighborhoods are at lower population densities and there is lower inequality in tree cover and surface temperature between neighborhoods³⁸.

National benefits of increased urban tree cover

Reforestation in our sample cities to the maximal reforestation scenario (bringing each block up to the 90th percentile observed in each impervious surface category in each city; see methods section for details) could reduce population-weighted mean summer air temperatures by 0.38 ± 0.014⁰C (Table 2), with reductions up to 1.8⁰C possible for specific neighborhoods. We estimate that a maximal reforestation scenario would involve the addition of 1.2 billion trees and would reduce annual heat-related mortality by 464 ± 83 people, additional to the current protection benefits that trees provide. There is roughly similar potential to reduce mortality in white and POC neighborhoods under the maximum planting scenario (Fig. 1).

The maximal reforestation scenario would deliver other benefits as well. It would reduce annual heat-related morbidity by an additional 80785 ± 6110 cases and would increase net carbon sequestration in trees by 23.7 ± 0.2 MtCO₂e/yr above the status quo (Table 2). It would also reduce electricity consumption by 4.3 ± 0.2 TWhr/yr, avoiding electricity related GHG emissions of 2.1 ± 0.1 MtCO₂e/yr. Note, however, that even lower levels of tree canopy cover increases could still deliver substantial benefits. For instance, aiming for a nominal 5% increase in urban tree canopy cover where feasible would reduce annual morbidity by 23614 ± 1832 cases (29% of the morbidity reduction achieved under the maximal planting scenario).

Table 2

**Estimated potential of urban reforestation to provide benefits to human well-being for increases in tree canopy cover in 5% increments above current tree canopy.** Shown are the annual reduction in mortality, morbidity, and electricity during heat waves, as well as the additional annual carbon storage. Error ranges show the 95% confidence interval of the estimate. Planting targets are evaluated within individual census blocks, and when it is not possible to realistically achieve a target on a census block the maximum possible planting on that block is modeled. The “Max” target scenario shows hitting the maximum possible planting on each block, across the 5,723 municipalities in our sample. See text for details.
Target	Reduction in median summer air temperature due to additional tree cover (⁰C)	Additional annual avoided mortality	Additional annual avoided morbidity	Additional annual avoided electricity consumption (TWhr)	Additional carbon storage (MtCO2e)
5%	0.11 ± 0.004	138 ± 25	23614 ± 1832	1.2 ± 0.1	5.6 ± 0.1
10%	0.19 ± 0.007	250 ± 44	42855 ± 3280	2.2 ± 0.1	10.4 ± 0.1
15%	0.26 ± 0.010	331 ± 59	56993 ± 4340	3.0 ± 0.2	14.2 ± 0.2
20%	0.31 ± 0.011	386 ± 69	66535 ± 5064	3.5 ± 0.2	17.1 ± 0.2
25%	0.34 ± 0.013	420 ± 75	72626 ± 5524	3.8 ± 0.2	19.2 ± 0.2
30%	0.36 ± 0.013	440 ± 79	76260 ± 5792	4.0 ± 0.2	20.6 ± 0.2
35%	0.37 ± 0.014	451 ± 81	78326 ± 5938	4.1 ± 0.2	21.7 ± 0.2
40%	0.37 ± 0.014	457 ± 82	79468 ± 6017	4.2 ± 0.2	22.4 ± 0.2
Max	0.38 ± 0.014	464 ± 83	80785 ± 6110	4.3 ± 0.2	23.7 ± 0.2

Spatial pattern of benefits of additional tree planting

Under the maximal reforestation scenario, the largest reforestation potential, in terms of trees planted or additional net carbon sequestration, occurs in the urbanized areas with the largest geographic extent (Fig. 4). For instance, the Chicago and New York City urban areas are among the largest in extent in the United States, and both have an increase in net carbon sequestration of more than 1 MtCO₂e/yr under the maximal reforestation scenario. However, the potential additional net carbon sequestration also varies by climate, with arid regions generally having lower feasible tree canopy targets due to natural limits on tree canopy in arid climates (see Methods for details), and humid regions in naturally forested biomes having higher feasible tree canopy targets. Similarly, the potential additional net carbon sequestration also varies by population density, with greater reforestation possible in lower population density urbanized areas, which have less impervious surface and hence higher proportions of plantable area. For the same reason, within urban areas, the greatest reforestation potential is often in suburban neighborhoods that are often majority white (e.g., Fig. 2, lower right). If urban reforestation was to be prioritized solely based upon the potential additional carbon sequestration, then suburban neighborhoods would often have priority.

However, patterns are different with regards to tree canopy cover’s role in reducing mortality and morbidity. The greatest increase in protective rate (avoided annual deaths per M people) under the maximal reforestation scenario occurs in the Northeastern and Western US (Fig. 4). These urbanized areas are relatively dense, with a high proportion of impervious surfaces. The addition of new tree canopy cover thus tends to shade impervious surfaces, reducing the urban heat island effect more substantially than if tree canopy shaded grass or other sparse vegetation. Moreover, these urbanized areas are in naturally forested biomes, so larger increases in tree cover are possible than in arid areas, where water limitations might limit tree canopy cover expansion.

Within urbanized areas, the greatest increase in protective rate under the maximal reforestation scenarios is in dense neighborhoods, often located in city centers and often majority POC neighborhoods (e.g., Fig. 2, lower right). Thus, if urban reforestation was to be prioritized solely based upon the potential protective health benefits of trees, then dense urban neighborhoods would have priority. For this reason, on average across the US, the return-on-investment (ROI) of tree planting in terms of reduction in annual heat-related mortality per tree planted is greater in majority POC neighborhoods than in majority white neighborhoods (Fig. 5, top panel). This is true for any level of planting ambition, from the 5% nominal target to the maximal reforestation scenario. For instance, the 5% nominal target for majority POC neighborhoods would avoid more mortality than the 5% nominal target for majority white neighborhoods, despite the number of trees required to achieve that target in POC neighborhoods being about half that needed in white neighborhoods. It is worth stressing, however, that there is significant variation among neighborhoods in ROI. If we examine mortality reduction benefits of the 5% of neighborhoods with the highest ROI (Fig. 5, bottom panel), we find the same general pattern of POC neighborhoods having higher ROI than white neighborhoods, but the number of deaths reduced per million trees planted is an order of magnitude greater in these high priority blocks than the overall national average.

Economic benefits of tree canopy increases

Total economic benefits of additional tree planting are shown in Table 3, for the 5% nominal target and the maximal reforestation scenario. Benefits valued in this study are avoided mortality, morbidity, electricity consumption, avoided GHG emissions from avoided electricity generation, and carbon sequestration. Under the maximal tree planting scenario, USD 9.6 ± 0.4 billion in annual benefits would accrue. The 5% target tree planting scenario delivers fewer annual benefits, USD 2.5 ± 0.1 billion. The largest economic value is in the carbon sequestered in the trees, followed closely by the avoided mortality due to heat. Avoided electricity consumption, avoided GHG emissions from electricity consumption, and the avoided morbidity due to heat all have meaningful but small total economic values.

Table 3

**Economic benefits of tree planting for 5,723 municipalities in the United States.** Shown are results for the 5% planting scenario and the maximum planting scenario.
	Economic value (Annual millions USD)
Variable	5% target	Max. target
Avoided mortality	982 ± 101	3299 ± 339
Avoided morbidity	32 ± 2	108 ± 7
Avoided electricity consumption	204 ± 14	703 ± 46
Carbon sequestration	1221 ± 65	5166 ± 287
Avoided GHG emissions from electricity	98 ± 6	344 ± 22
TOTAL	2536 ± 121	9620 ± 447

The economic benefits of the maximal planting scenario are split relatively equally (Table 4) between POC neighborhoods (USD 3.9 ± 0.2 billion) and white neighborhoods (USD 5.0 ± 0.2 billion). However, annual tree planting and maintenance costs would exceed annual benefits for both POC neighborhoods (costs of USD 19.5 ± 4.4 billion, Return on Investment, ROI = 0.20) and, especially, white neighborhoods (USD 40.5 ± 9.4 billion, ROI = 0.12). However, if tree canopy increases were targeted to the neighborhoods with the highest ROI, the situation looks different, with the annual economic benefits roughly equal to the costs of annual tree planting and maintenance. Indeed, a 5% nominal target for additional tree planting in High-ROI POC neighborhoods is estimated to provide USD 32 ± 4 million in benefits but to cost only USD 29 ± 7 million in planting and maintenance, an ROI of 1.12.

Table 4

**Economic benefits and costs of additional tree planting, by race/ethnicity of census blocks.** Shown is data for tree planting only in populated census blocks where human well-being benefits occur. Data are shown for neighborhoods that are majority people of color (POC) as well as majority non-Hispanic white (White). Census blocks are split into two priority categories, high return on investment (ROI) blocks and other blocks. High ROI is defined as neighborhoods in the top 5% of ROI.
		Economic benefits (annual million USD)		Cost of planting and maintenance (annual million USD)		ROI (benefits/costs)
Target	Priority	POC	White	POC	White	POC	White
5% increase in tree cover	High ROI	32 ± 4	18 ± 2	29 ± 7	19 ± 4	1.12	0.95
5% increase in tree cover	Other blocks	1053 ± 64	1283 ± 49	4696 ± 1084	9794 ± 2261	0.22	0.13
	All blocks	1086 ± 64	1301 ± 49	4725 ± 1084	9813 ± 2261	0.23	0.13
Max. increase in tree cover	High ROI	81 ± 10	45 ± 5	86 ± 20	58 ± 13	0.94	0.77
Max. increase in tree cover	Other blocks	3805 ± 222	4967 ± 193	19464 ± 4493	40517 ± 9352	0.20	0.12
	All blocks	3886 ± 222	5012 ± 193	19550 ± 4493	40575 ± 9352	0.20	0.12

Our results show that trees provide substantial benefits in the US, but that inequality in urban tree canopy cover leads to inequality in the protective benefits that trees provide against heat-related mortality and morbidity. In our sample of 5,723 municipalities, we estimate that, annually, trees in white neighborhoods help avoid 190 more deaths, 30,131 more doctors’ visits, and 1.4 TWhr more electricity consumption than in POC neighborhoods, despite the nearly equal number of people in white and POC neighborhoods. While POC neighborhoods have lower tree cover than white neighborhoods, their higher impervious surface area means that tree cover is more often shading impervious surfaces and thus provides relatively large cooling benefits per unit canopy area. Climate change is expected to increase the frequency and intensity of heat waves, increasing mortality and morbidity⁷. This will likely elevate the public health importance of heat action planning in the coming decades. We believe that heat action planning should consider outdoor temperature, its public health impacts, the equity of these impacts, and strategies to mitigate these impacts and their inequity, such as increased investment in tree planting and maintenance.

We found that the maximum feasible urban reforestation scenario in these 5,723 municipalities would annually save an additional 464 lives otherwise threatened by heat, equal to around 8% of current total heat related deaths in the US³. In comparison, a recent paper in Europe estimated that tree planting up to a 30% target in all neighborhoods would reduce Europe deaths from heat by around 33%⁵⁸. One potential reason for the difference between studies is the differences in how targets were set. Our targets considered physical and social limitations in dense neighborhoods to tree planting while the European study did not. Another potential reason is simply that European cities are on average denser with more impervious surface than American cities, so tree cover is more likely to shade impervious surfaces resulting in more significant heat reduction benefits.

Trees provide other benefits as well. We estimated that the maximum feasible urban restoration scenario could reduce annual heat-related morbidity by an additional 80,785 cases, increase carbon sequestration in trees by 23.7 MtCO2e/yr, and decrease electricity-related GHG emissions by 2.1 ± 0.1 MtCO2e/yr. While expanding the urban forest by the additional 1.2 billion trees needed under this maximal scenario would be a large financial investment, we want to emphasize that the heat-health benefits would likely offset a significant fraction of the costs. Indeed, targeted actions to increase tree canopy in high-ROI neighborhoods can result in benefits that equal or exceed tree planting and maintenance costs. Trees, of course, provide many other benefits not modelled in this paper, both to overall health⁵⁹ and to human well-being⁴⁵, and consideration of these other benefits would increase the total estimated benefits and ROI of increased urban tree canopy.

There is greater return-on-investment of tree cover increases in majority POC neighborhoods for health benefits, relative to majority white neighborhoods. That is, for a given number of trees planted, urban greening projects in POC neighborhoods will, on average, have a greater reduction in mortality and morbidity than projects in majority white neighborhoods. Similarly, tree planting projects that target neighborhoods with disproportionately low tree cover will have a disproportionate benefit to POC households, because POC households tend to be in neighborhoods with lower tree cover. Given the high tree cover inequality in the U.S. with respect to race/ethnicity and income³⁸, programs such as the Inflation Reduction Act that aim to reduce heat-health impacts must address tree cover inequality, either explicitly or implicitly, in order to target trees where they will have the greatest benefit. Of course, there are other reasons to consider race/ethnicity and socioeconomic status when doing heat action planning, as these populations are more vulnerable to heat hazards due to, among other things, lower access to air conditioning⁶⁰.

Our estimated maximum carbon sequestration potential of urban reforestation in our sample of 5,723 municipalities, 23.7 MtCO₂e/yr, is lower than Cook-Patton et al.⁴⁹ of 52.5 Mt CO₂/yr and higher than that of Fargione et al.⁴⁴ of 23.3 Mt CO₂/yr, both of whom were estimating for all urban areas. Another difference is that our definition of what could potentially be planted is more restrictive than in Cook-Patton et al.⁴⁹, due to our consideration of both physical and social barriers to planting, which may further explain why our estimate is lower. Our estimation of urban reforestation potential for sequestration is about 2.6% of the overall U.S. potential for NCS across all pathways in the US identified by Fargione et al.⁴⁴. For context, urbanized areas in the US occupy about 2.8% of the total land area of the US⁶¹. We also estimated that avoided electricity consumption due to additional trees would decrease electricity-related GHG emissions by 2.1 MtCO₂e/yr, for a total climate mitigation potential of 25.8 MtCO₂e/yr, a modest amount of carbon relative to the entire U.S. carbon budget, but not insignificant, being the annual equivalent emissions of 5.7 million gasoline-powered cars⁶².

We caution that there is, to some extent, a spatial disconnect between optimal sites for climate mitigation and climate adaptation. Although the carbon benefit per tree planted did not vary in our study and is expected to be the same from exurbs to downtown, the overall potential for carbon sequestration is greater in neighborhoods with more plantable area (i.e., fewer impervious surfaces). This is more frequently the case in the suburbs or exurbs than in the city center. Patterns for avoided electricity consumption (and hence avoided GHG emissions) are the opposite, being greater in densely populated blocks, but are an order of magnitude smaller than carbon sequestration in terms of Mt CO₂/yr. Heat-related health benefits also are greatest when trees are planted close to where people live and work, which is more frequently the case in city centers. We note that greater heat-related mitigation benefits in more densely populated neighborhoods would likely hold for other tree-provided ecosystem services that need to be generated close to people, such as air pollution mitigation, stormwater mitigation, and the benefits to physical and mental health from exposure to nature⁴⁵. The varying spatial scales of different ecosystem services lead to different optimal places of tree planting⁶³, with heat-reduction benefits needing to be provided within 100m or so of people⁵. Carbon sequestration is, on the other hand, essentially global, since the atmosphere is well-mixed, and not restricted to locations near people. We stress however that this spatial disconnect between heat mitigation and carbon mitigation benefits is only partial, and sites can be found that have good returns for both. If both benefits are important, planners would do well to consider the joint return-on-investments from both benefits when making decisions about where to plant.

Realizing the climate mitigation and adaptation benefits of the urban reforestation scenario considered in this paper requires overcoming the “wrong pocket problem”⁶⁴. This problem occurs when the person or institution that pays for an action is not the one benefiting from that action. Urban forestry agencies in municipalities do not generally have climate mitigation or adaptation as part of their mission, so while they must pay for increased tree planting and maintenance, they may not see a benefit in terms of their mission. Conversely, health agencies see heat action planning as part of their mission and would welcome the mortality and morbidity reductions of increased tree planting, but many not have capacity to plant trees themselves. Regulations, incentives, or investments created to increase the provision of ecosystem services can give value to these benefits and solve the wrong pocket problem by creating an incentive for different institutions to work together.

Investment at a higher level of government by agencies who have as part of their mission climate mitigation or climate adaptation can overcome the sometimes-narrower goals and capacities of municipal forestry offices. To the extent natural climate solutions are considered as part of federal policies to mitigate and adapt to climate change, urban reforestation may be an investment that generates relatively modest carbon sequestration but delivers significant local climate adaptation benefits⁶⁵. This may be particularly important in low-income neighborhoods, which generally have lower tree cover and are hotter than high-income neighborhoods^38,66. Although there is a growing interest in urban reforestation to meet climate mitigation goals, our work shows that a single-minded focus on carbon sequestration could further exacerbate inequities by shifting planting efforts outside of historically people-of-color neighborhoods. Instead, putting equity first and making targeted investments in tree planting and maintenance in neighborhoods most at risk from heat could help correct some of this historic inequality in tree cover with respect to race.

Our analysis proceeded in four phases. First, we assembled spatial data from multiple sources and compiled them to a common analysis unit. Second, we developed an algorithm that would find the maximum urban reforestation realistically possible, given other land-use constraints. Third, we estimated the heat mitigation-related benefits of current tree canopy and of future planting scenarios, up to the maximum urban reforestation realistically possible. Benefits evaluated were avoided mortality, avoided morbidity, avoided electricity consumption, avoided release of greenhouse gases from avoided electricity consumption, and carbon sequestration in aboveground tree biomass. Fourth, we valued these benefits in monetary terms. In the following section, we discuss each of these phases in detail.

Data sources

Our study focused on all US urbanized areas larger than 500 km² in extent³⁸. Urbanized area is defined by the US Census Bureau as contiguous census blocks that exceed a population density threshold as well as adjacent territory that links nearby patches of high-density settlements⁶¹. There are 100 urbanized areas that are greater than 500 km² in extent, housing 180 million people during the 2020 census. Each urbanized area contains many municipalities or other census-designated places; in total, our study area contains 5,723 such communities³⁸. These span the range of population sizes and densities, from a large central city like New York City (8.4 million people) to small communities at the edges of metro areas with just a few thousand people. These urbanized areas also span the range of biomes in the United States, from forests to deserts to grasslands.

Our tree cover maps are taken from McDonald et al.³⁸, which mapped tree cover for the same set of 100 urbanized areas, using the boundaries of the urbanized area set after the 2010 Census. For consistency with that dataset, we use the same boundary file for this analysis, rather than the 2020 urbanized area boundaries. Aerial photos taken during the summer growing season at approximately 2m resolution from the National Agricultural Imagery Program (NAIP) were classified into a forest/non-forest grid using Google Earth Engine’s cloud computing platform. Training data for the supervised classification came from Nowak and Greenfield⁶⁷, which examined tree cover at 10,000 control points throughout urban and community areas of the United States. After creating several texture variables and band indices (e.g., NDVI), this information, along with information from the original spectral bands, was classified using a random forests classifier built off the training data. Pixel-level classification accuracy varies by biome but averages around 75%. In this study, we use tree cover estimates aggregated up to the US Census block level. Estimates at this scale were shown to be extremely accurate (R² = 0.94) compared to an external validation dataset. Much more info on our classification methodology is available in McDonald et al.³⁸.

Our information on demography and socio-economic status comes from the US Census Bureau’s 2020 decennial census, downloaded from National Historical Geographic Information System (NHGIS)⁶⁸. Population and its breakdown by race and ethnicity were available at the Census block level. Details of the GIS analysis and processing of this data can be found in McDonald et al.³⁸, which used the 2010 Census but otherwise had similar processing of Census data.

Impervious surface data were taken from the National Land Cover Database 2019 product⁶⁹. This dataset was derived from Landsat imagery, and measures as a continuous variable the percent impervious surface in each pixel. While 30m data on impervious surface cover is coarser than our forest cover map, the fact that there is a continuous estimate of imperviousness helps address sub-pixel heterogeneity.

Land surface temperature (LST) was derived from the Collection 2 LST science product from Landsat 8 at 100 m resolution. Pixel level quality flags were used to remove cloud contaminated pixels, following the methodology of Chakraborty et al. 2022⁵⁰. All data between 2016 and 2020 were accessed through the Google Earth Engine platform⁷⁰ to create mean summer (June, July, August; JJA) LST composites. This 5-year period averages over year-to-year variability in LST, while also being long enough to ensure that most regions have usable, cloud-free pixels for which LST can be estimated.

Maximum urban reforestation and tree planting scenarios

Next, we defined an algorithm to find the maximum urban reforestation realistically possible, given other land-use and climatic constraints. Conceptually, we wanted to consider two constraints: a physical constraint (trees can generally only be planted in non-impervious surfaces, excluding rare landscape features such as planter boxes) and a social/political/climatic constraint (planting trees is constrained by numerous other considerations, including competing land uses, landowner preferences, zoning and building codes, and climatic conditions).

Tree cover is negatively correlated with impervious surface cover³⁸. We divided the landscape into five impervious surface categories (0–20%, 20–40%, 40–60%, 60–80%, 80–100%) to reflect the different landscape contexts from lightly settled suburbs to dense urban core. Our general strategy was to model different scenarios, in each US Census block, where tree cover was increased by intervals of 5% (e.g., from 5–10%) potentially up to the maximum possible where all non-impervious land surfaces were greened. However, tree planting was stopped when there were social/political constraints. Social/political/climatic constraints were defined at the 90th percentile of observed tree cover in each impervious surface category for each urbanized area. In other words, tree planting was halted at the 90th percentile of tree cover for sites in that urbanized area with similar landscape contexts (imperviousness). This approach implicitly accounts for differences among cities (e.g., in climate), as well as urban form and zoning and building codes.

Once the maximum possible urban reforestation was estimated at the Census block level, we constructed a set of scenarios for all our sample cities, each aiming to hit a certain nominal increase in tree cover. So, for instance, the 5% increase planting scenario aimed to increase tree cover in each census block by 5%. If this nominal increase exceeded the maximum possible urban tree cover increase for a census block, then tree cover for that census block was set at the maximum possible urban tree cover amount instead. We constructed planting scenarios with increases of 5%, 10%, 15%, 20%, 25%, 30%, 35%, 40%, and Maximum (i.e., all census blocks set at the maximum possible urban reforestation amount), respectively.

Note that an increase in tree canopy cover, as seen from directly overhead a city and above the tree canopy, necessarily decreases other land covers, since total land cover in a region must sum to 100%. To account for this in our planting scenarios, we made the simplest possible assumption, that additional tree cover occurs randomly over other land uses, in proportion to their cover before planting. For instance, if in an impervious surface category in a particular city, 25% of the area not already in tree cover was impervious and 75% was in sparse vegetation, we assumed that any tree canopy increases would occur 25% over impervious surfaces and 75% over sparse vegetation, displacing these land cover types accordingly.

Finally, we calculated the number of new stems that would need to be planted to achieve this increase in tree canopy, based on the average canopy size of 19.6 m²/stem derived from Nowak and Greenfield⁷¹. Note that planted trees grow their tree canopy over time, with some tree species in mesic environments taking 3–5 years to reach 5m height and diameter and more than 10 years to reach 10m height and diameter⁷². Our tree planting scenarios assume spacing appropriate for adult trees, and we are estimating benefits for the tree canopy associated with those adult trees.

Estimating benefits

Values of average annual net carbon sequestration per m² of urban tree canopy were taken from Nowak et al.⁷³. They assembled data on tree growth rates and carbon accumulation in tree biomass and subtracted out GHG emissions during planting and maintenance to estimate there was 0.226 kg C annual average net carbon sequestration per m² of canopy.

To estimate the likely reduction in summer air temperature that would occur due to urban reforestation, we used a two-step regression approach, similar to that of Zhang et al.⁷⁴. First, we statistically related impervious cover and urban tree cover to LST at the census block level. Urbanized areas were divided into two climatic groups, mesic and xeric, based on biomes (see McDonald et al. ²⁷ for details). For each climatic group, a linear regression was conducted using PROC GLM in SAS, relating LST (the dependent variable) to both impervious cover and urban tree cover. A fixed effect for urbanized area was also included. To avoid potential issues of spatial autocorrelation, we take a sparse sample of 10,000 census blocks out of the 2.3 million census blocks in our sample area. The sparse sample represents 0.4% of the total number of Census blocks and ensures that census blocks are on average quite far from one another and are thus more likely to be independent statistically in terms of the dependent variable, LST (i.e., any errors in estimation of LST in one census block are likely to be spatially uncorrelated with errors in other census blocks). Five observations were dropped due to missing data. Results from our regression are shown in Table 5. The R² for xeric regions was 0.85, while that for mesic regions was 0.76.

Table 5

**Model parameters from regression relating LST to tree cover and impervious surface cover.** Separate regression analyses were undertaken for mesic and arid cities. A fixed effect for urbanized area (NAME) was included.
Census blocks in mesic cities (N = 8498 with complete data)
Variable	Type III SS	Mean SS	F value	P	Estimate	SE
Tree cover (fraction)	1443	1443	310	< 0.001	-2.88	0.16
Impervious surface cover (fraction)	15684	15684	3367	< 0.001	9.06	0.16
NAME (N = 86)	68869	810	174	< 0.001	multiple
Census blocks in arid cities (N = 1497 with complete data)
Variable	Type III SS	Mean SS	F value	P	Estimate	SE
Tree cover (fraction)	1031	1031	212	< 0.001	-8.06	0.55
Impervious surface cover (fraction)	327	327	67	< 0.001	2.78	0.34
NAME (N = 14)	35906	2762	567	< 0.001	multiple

Second, we statistically related air temperature information with LST. The air temperature data came from the Global Historical Climatology Network⁷⁵ for summer (JJA) average daily mean (see McDonald et al.²⁷ for details on data handling and processing). There were 423 air temperature stations within our study area. A linear regression was conducted using PROC GLM in SAS, relating air temperature (the dependent variable) to LST. The fitted slope was allowed to vary by biome (i.e., biome is included as a fixed effect in interaction with LST in the model). A fixed effect for urbanized area was also included. The number of air temperature sensors is relatively small, they are measured independently by different machines, and they are relatively far from one another, so we did not suspect there would be spatial autocorrelation in the dependent variable after accounting for LST. Results from our regression are shown in Table 6. The R² for this regression was 0.49, lower than for our statistical model explaining LST, perhaps reflecting the fact that there is a broad variety of factors that affect air temperature beyond the local surface temperature⁷⁶. Regardless, we can propagate any uncertainty in estimating air temperature in our analysis, which is an advantage of our statistical approach as opposed to more mechanistic models.

Table 6

**Model parameters from regression relating air temperature to LST.** Regression slopes were allowed to vary by the biome in which an urbanized area is located. The overall regression also included an intercept term, and was highly significant (F = 57.8, P < 0.001).
Census blocks with air sensors (N = 423)
Biomes	Estimate	SE
Temperate Broadleaf and Mixed Forests	0.33	0.030
Temperate Coniferous Forests	0.35	0.030
Tropical and subtropical grasslands, savannas, and shrublands	0.40	0.033
Temperate Grasslands, Savannas, and Shrublands	0.36	0.028
Flooded Grasslands and Savannas	0.46	0.034
Mediterranean Forests, Woodlands, and Scrub	0.25	0.026
Deserts and Xeric Shrublands	0.35	0.023

Note that therefore, the air temperature impacts of our tree planting scenarios are a function of multiple factors: the amount of plantable area in a Census block; the fraction of impervious surface before planting, which determines how much impervious surface cover is reduced through planting; the aridity and biome of an urbanized area, and an urbanized area fixed effect. We summarize the effect of multiple factors in Table S1, which lists (by urbanized area and impervious surface category) the realized change in tree cover and impervious surfaces for the 5% target scenario, as well as the estimated change in LST and air temperature. Not surprisingly, the greatest decline in impervious surface area with planting is generally in impervious surface category 5 (defined as neighborhoods with 80–100% impervious surface cover). The only exception is when there is not enough plantable area in impervious category 5 neighborhoods to approach the 5% target for tree canopy increase. Conversely, the smallest decline in impervious surface area with planting is generally in impervious surface category 1 (defined as neighborhoods with 0–20% impervious surface cover). For this reason, the greatest decline in LST occurs in impervious category 5 while there are lesser declines in impervious category 1, with similar patterns across climate variables (aridity and biome). Asheville, NC, has the greatest decline in LST in impervious category 5 (0.49 ⁰C). Patterns for air temperature declines are similarly generally greater in impervious category 5 and less in impervious category 1, with similar patterns across climate variables (aridity and biome). Bonita Spring, FL, has the greatest decline in air temperature in category 5 (0.21 ⁰C). Finally, note that we estimate statistically the change in LST and air temperature at the scale of a US census block, which includes a variety of land-use types. This makes our estimates of temperature impact per unit increase in tree cover different from (and lower in magnitude than) the estimates of tree cooling efficiency of Zhao et al.⁷⁷, who used Landsat-based estimates of tree cover and LST to quantify change in LST for those pixels with tree canopy.

This study focuses on tree canopy and its benefits relative to the race/ethnicity of neighborhoods. Within our sample of cities, race/ethnicity is correlated to the impervious surface category (Table 7), with most neighborhoods in categories 4 and 5 being majority POC, while most neighborhoods in categories 1 and 2 are white. This trend is discussed extensively in McDonald et al.³⁸, and has to do with low-income households, which are more likely to be POC, being more frequently located in US Census blocks with high population density and high impervious surface cover in city centers rather than less dense suburbs. Race/ethnicity is also correlated with aridity, as the arid southwest of the US has a larger fraction of POC in all impervious categories than does the mesic region of the country. This trend primarily has to do with people of Hispanic origin, who are counted in the POC category for our study. Since impervious surface category and aridity are two of the explanatory variables used to predict the impact of tree canopy increases on air temperature, the correlation of race/ethnicity with these variables necessarily implies that the impact of tree canopy increases on air temperature varies with respect to race/ethnicity.

Table 7

**Race/ethnicity composition of our sample, relative to impervious surface and aridity.** Population totals, in millions, are shown. Percentage of population within each aridity category and each impervious surface category are shown in parentheses. For instance, in arid urbanized areas, there are 2.7 million POC in impervious surface category 5, 78% of the total population of arid cities in this impervious surface category.
	Arid population, millions (%)		Mesic population, millions (%)
Impervious surface category	White	POC	White	POC	Total population, millions
1 (0–20%)	0.7 (54%)	0.6 (46%)	15.6 (79%)	4.1 (21%)	21.0
2 (20–40%)	2.0 (54%)	1.6 (46%)	25.9 (66%)	13.4 (34%)	42.9
3 (40–60%)	4.5 (36%)	8.1 (64%)	22.9 (52%)	20.7 (48%)	56.2
4 (60–80%)	3.0 (22%)	10.8 (78%)	10.6 (38%)	17.2 (62%)	41.8
5 (80–100%)	0.8 (22%)	2.7 (78%)	5.2 (36%)	9.2 (64%)	17.8
Overall categories	(32%)	(68%)	(55%)	(45%)

To estimate the health impacts of an estimated change in air temperature due to tree canopy expansion, we follow the methodology of McDonald et al.²⁷. For estimating avoided mortality, we use published epidemiological studies that relate changes in air temperatures to changes in mortality over time. Specifically, we use Bobb et al.⁷⁸, who estimated heat-mortality relationship for 105 US cities. In this study, we use Bobb et al.’s regional estimates of the heat-mortality response curve. For morbidity estimation, our analysis is based on Gronlund et al.⁷⁹, who estimated heat-related hospitalizations as a function of air temperature for a large population in the US, and then scaled from the Gronlund et al. numbers to estimate emergency department visits and doctor’s office visits²⁷.

We estimated avoided electricity consumption due to tree cover, following the methodology in McDonald et al.²⁷. We base our analysis off Santamouris et al. ⁴¹, a literature review that collected empirical estimates of increases in electricity use during periods of high air temperature. We subset the results in Table 1 of their paper to just look at US studies, which gave a range of 2.9–8.5% increase in electricity consumption per 1°C increase in air temperature. We assume that the shift in summer daily mean temperatures we modeled would only increase electricity consumption for space cooling during summer months (1/4 of the year). From this assumption, we calculated an estimated annual increase in electricity consumption, assuming no increase in other seasons. We then multiplied percent increase in electricity consumption by city-level data on residential electricity use. To estimate avoided GHG emissions due to avoided electricity consumption, we multiplied avoided electricity consumption by the average carbon intensity of US electricity generation, taken from EPA data.

Finally, when calculating statistics at the level of urbanized areas or nationally, we used population-weighted statistics, to give greater weight to Census Blocks with greater population.

Valuation and extrapolation

Our valuation methodology follows McDonald et al.²⁷, and is extensively described in the supplementary methods section of that paper. To briefly summarize, for avoided mortality we use a value of a statistical life approach (VSL), adjusted for remaining life years e.g., ^80,81, using a range in the US of $5.4-13.4M (2015$)⁸². Most heat-related deaths occur in people 55 years or older^80,83, and so we apply Murphy and Topel’s⁸¹ life cycle shape for the value of a life-year for 11 age classes to calculate an overall average, age-weighted VSL, accounting for the age distribution of mortality from heat-related events, of $5.7M (2015$, range: $3.3-8.2M).

For morbidity, we use a cost-of-illness (COI) approach to assess the economic burden associated with heat-related illnesses (HRI) that were avoided due to tree cover. Economic calculations were done separately for avoided emergency department (ED) visits⁸⁴, avoided hospitalizations⁸⁵, and outpatient visits⁸⁶. We also estimated lost work productivity as the product of total work loss days (from HRI-related hospitalizations, ED visits and outpatient visits) and the average US daily earnings rate (see McDonald et al.²⁷ for details).

For electricity consumption, data on average household residential electricity consumption and average cost per KWh, by electric utility, were taken from the US Energy Information Administration (EIA, 2016) form EIA-861, submitted by electric utilities annually.

For carbon sequestration and avoided GHG emissions due to avoided electricity consumption, we used the new EPA proposed Social Cost of Carbon (SCC) for 2020 of USD 190 expressed at 2020 prices, assuming a 2% discount rate⁸⁷. This estimate follows that of recent science by Rennert et al.⁸⁸. The SCC measures the total (global) estimated economic damages, expressed in monetary terms, that result from one additional ton of carbon dioxide in the atmosphere. These damages from increased atmospheric CO₂ concentrations are completely separate from, and fully additional to, the local damages from health and electricity use impacts that increases in urban tree canopy avoid via their local cooling effects. Where necessary, all economic values presented in this paper were standardized to USD in 2022, using the US Bureau of Labor Statistics Consumer Price Index⁸⁹.

Author contributions: RM designed and led the analysis. TB calculated tree cover for the municipalities in our sample. TC calculated land-surface temperatures. TK designed the economic valuation analysis. SC-P helped design the rules for where reforestation was feasible, and with JF helped design the calculation of carbon sequestration benefits. All co-authors helped write the manuscript.

Acknowledgements: TC was supported by the Pacific Northwest National Laboratory, which is operated for DOE by Battelle Memorial Institute under contract DE-AC05-76RL01830, and by COMPASS-GLM, a multi-institutional project supported by the U.S. DOE, Office of Science. Co-authors who work at The Nature Conservancy (TNC) were supported by the members and donors of TNC. The funders played no role in study design, data collection, analysis and interpretation of data, or the writing of this manuscript.

Competing Interests: All authors declare no financial or non-financial competing interests.

Data availability: The datasets generated and/or analysed during the current study are available in the Data Dryad repository, [during review this data is not yet available online, if reviewers wish to access it, please ask the journal editor and the authors can make it available].

Code availability: The underlying code for this study is available in Zenodo and can be accessed via this link [during review this code is not yet available online, if reviewers wish to access it, please ask the journal editor and the authors can make it available].

Hsiang, S. et al. Estimating economic damage from climate change in the United States. Science 356, 1362–1369 (2017).
Neumann, J. E. et al. Climate damage functions for estimating the economic impacts of climate change in the United States. Review of environmental economics and policy (2020).
Burkart, K. G. et al. Estimating the cause-specific relative risks of non-optimal temperature on daily mortality: a two-part modelling approach applied to the Global Burden of Disease Study. The Lancet 398, 685–697, doi:10.1016/S0140-6736(21)01700-1 (2021).
McMichael, A. et al. in Comparative Quantification of Health Risks: Global and regional burden of disease attributable to selected major risk factors (eds Majid Ezzati, Alan D. Lopez, A. Rodgers, & C. Murray) (World Health Organization, 2004).
McDonald, R. I., Kroeger, T., Boucher, T., Wang, L. & Salem, R. Planting Healthy Air: A global analysis of the role of urban trees in addressing particulate matter pollution and extreme heat. (The Nature Conservancy, Arlington, VA, 2016).
Spanger-Siegfried, E. et al. Killer Heat in the United States: Climate Choices and the Future of Dangerously Hot Days. (Union of Concerned Scientists, 2019).
Gasparrini, A. et al. Projections of temperature-related excess mortality under climate change scenarios. The Lancet Planetary Health 1, e360-e367 (2017).
Limaye, V. S., Vargo, J., Harkey, M., Holloway, T. & Patz, J. A. Climate change and heat-related excess mortality in the Eastern USA. EcoHealth 15, 485–496 (2018).
Reidmiller, D. R. et al. Impacts, risks, and adaptation in the United States: Fourth national climate assessment, volume II. (2017).
Luber, G. & McGeehin, M. Climate change and extreme heat events. American journal of preventive medicine 35, 429–435 (2008).
Oke, T. R. The Energetic Basis of the Urban Heat Island. Quarterly Journal of the Royal Meteorological Society 108, 1–24 (1982).
Ziter, C. D., Pedersen, E. J., Kucharik, C. J. & Turner, M. G. Scale-dependent interactions between tree canopy cover and impervious surfaces reduce daytime urban heat during summer. Proceedings of the National Academy of Sciences 116, 7575–7580 (2019).
Chakraborty, T., Venter, Z. S., Qian, Y. & Lee, X. Lower Urban Humidity Moderates Outdoor Heat Stress. AGU Advances 3, e2022AV000729, doi:https://doi.org/10.1029/2022AV000729 (2022).
Piil, J. F. et al. Direct exposure of the head to solar heat radiation impairs motor-cognitive performance. Scientific reports 10, 7812 (2020).
Foster, J. et al. Quantifying the impact of heat on human physical work capacity; part II: the observed interaction of air velocity with temperature, humidity, sweat rate, and clothing is not captured by most heat stress indices. Int J Biometeorol 66, 507–520 (2022).
Matthies, F., Bickler, G., Marin, N. & Hales, S. (WHO Regional Office for Europe, Copenhagen, 2008).
Boeckmann, M. & Rohn, I. Is planned adaptation to heat reducing heat-related mortality and illness? A systematic review. BMC public health 14, 1112 (2014).
Schneider, F. A., Ortiz, J. C., Vanos, J. K., Sailor, D. J. & Middel, A. Evidence-based guidance on reflective pavement for urban heat mitigation in Arizona. Nature Communications 14, 1467 (2023).
Tan, H., Kotamarthi, R., Wang, J., Qian, Y. & Chakraborty, T. Impact of different roofing mitigation strategies on near-surface temperature and energy consumption over the Chicago metropolitan area during a heatwave event. Sci Total Environ 860, 160508 (2023).
Stone Jr, B. et al. Avoided heat-related mortality through climate adaptation strategies in three US cities. PLoS ONE 9, e100852 (2014).
Kroeger, T., McDonald, R. I., Boucher, T., Zhang, P. & Wang, L. Where the people are: Current trends and future potential targeted investments in urban trees for PM10 and temperature mitigation in 27 U.S. cities. Landsc Urban Plan 177, 227–240 (2018).
Declet-Barreto, J., Knowlton, K., Jenerette, G. D. & Buyantuev, A. Effects of urban vegetation on mitigating exposure of vulnerable populations to excessive heat in Cleveland, Ohio. Weather, Climate, and Society 8, 507–524 (2016).
Jenerette, G. D. et al. Micro-scale urban surface temperatures are related to land-cover features and residential heat related health impacts in Phoenix, AZ USA. Landsc Ecol 31, 745–760 (2016).
Middel, A., Chhetri, N. & Quay, R. Urban forestry and cool roofs: Assessment of heat mitigation strategies in Phoenix residential neighborhoods. Urban Forestry & Urban Greening 14, 178–186 (2015).
Coutts, A. M. et al. Thermal infrared remote sensing of urban heat: Hotspots, vegetation, and an assessment of techniques for use in urban planning. Remote Sens Environ 186, 637–651 (2016).
Middel, A., AlKhaled, S., Schneider, F. A., Hagen, B. & Coseo, P. 50 grades of shade. Bulletin of the American Meteorological Society 102, E1805-E1820 (2021).
McDonald, R. I., Kroeger, T., Zhang, P. & Hamel, P. The Value of US Urban Tree Cover for Reducing Heat-Related Health Impacts and Electricity Consumption. Ecosystems, 1–14 (2019).
Watkins, S. L. & Gerrish, E. The relationship between urban forests and race: A meta-analysis. J Environ Manage 209, 152–168 (2018).
Gerrish, E. & Watkins, S. L. The relationship between urban forests and income: A meta-analysis. Landsc Urban Plan 170, 293–308 (2018).
Duncan, D. T. et al. A spatially explicit approach to the study of socio-demographic inequality in the spatial distribution of trees across Boston neighborhoods. Spatial demography 2, 1–29 (2014).
Venter, Z. S., Shackleton, C. M., Van Staden, F., Selomane, O. & Masterson, V. A. Green Apartheid: Urban green infrastructure remains unequally distributed across income and race geographies in South Africa. Landsc Urban Plan 203, 103889 (2020).
Riley, C. B. & Gardiner, M. M. Examining the distributional equity of urban tree canopy cover and ecosystem services across United States cities. PLoS ONE 15, e0228499 (2020).
Schwarz, K. et al. Trees grow on money: urban tree canopy cover and environmental justice. PLoS ONE 10, e0122051 (2015).
Nesbitt, L., Meitner, M. J., Girling, C., Sheppard, S. R. & Lu, Y. Who has access to urban vegetation? A spatial analysis of distributional green equity in 10 US cities. Landsc Urban Plan 181, 51–79 (2019).
Casey, J. A., James, P., Cushing, L., Jesdale, B. M. & Morello-Frosch, R. Race, ethnicity, income concentration and 10-year change in urban greenness in the United States. International journal of environmental research and public health 14, 1546 (2017).
Corbane, C. et al. The grey-green divide: multi-temporal analysis of greenness across 10,000 urban centres derived from the Global Human Settlement Layer (GHSL). International Journal of Digital Earth 13, 101–118 (2020).
Jesdale, B. M., Morello-Frosch, R. & Cushing, L. The racial/ethnic distribution of heat risk–related land cover in relation to residential segregation. Environmental health perspectives 121, 811–817 (2013).
McDonald, R. I. et al. The tree cover and temperature disparity in US urbanized areas: Quantifying the association with income across 5,723 communities. PLoS ONE 16, e0249715 (2021).
Hsu, A., Sheriff, G., Chakraborty, T. & Manya, D. Disproportionate exposure to urban heat island intensity across major US cities. Nature communications 12, 2721 (2021).
Benz, S. A. & Burney, J. A. Widespread race and class disparities in surface urban heat extremes across the United States. Earth's Future 9, e2021EF002016 (2021).
Santamouris, M., Cartalis, C., Synnefa, A. & Kolokotsa, D. On the impact of urban heat island and global warming on the power demand and electricity consumption of buildings—A review. Energy and Buildings 98, 119–124 (2015).
EPA. Reducing Urban Heat Islands: Compendium of Strategies (draft). (Environmental Protection Agency, 2014).
Ko, Y. Trees and vegetation for residential energy conservation: A critical review for evidence-based urban greening in North America. Urban Forestry & Urban Greening 34, 318–335 (2018).
Fargione, J. E. et al. Natural climate solutions for the United States. Science Advances 4, eaat1869, doi:10.1126/sciadv.aat1869 (2018).
Keeler, B. L. et al. Social-ecological and technological factors moderate the value of urban nature. Nature Sustainability 2, 29–38, doi:10.1038/s41893-018-0202-1 (2019).
Griscom, B. W. et al. Natural climate solutions. Proceedings of the National Academy of Sciences 114, 11645–11650 (2017).
Nowak, D. J., Stevens, J. C., Sisinni, S. M. & Luley, C. J. Effects of urban tree management and species selection on atmospheric carbon dioxide. (2002).
Strohbach, M. W., Arnold, E. & Haase, D. The carbon footprint of urban green space—A life cycle approach. Landsc Urban Plan 104, 220–229 (2012).
Cook-Patton, S. et al. Lower cost and more feasible options to restore forest cover in the contiguous United States for climate mitigation. One Earth 3, 739–752 (2020).
Chakraborty, T. et al. Feasibility of afforestation as an equitable nature-based solution in urban areas. Sustainable Cities and Society 81, 103826 (2022).
Nowak, D. J. & Greenfield, E. J. Evaluating the National Land Cover Database tree canopy and impervious cover estimates across the conterminous United States: a comparison with photo-interpreted estimates. Environ Manage 46, 378–390 (2010).
Smith, M. L., Zhou, W., Cadenasso, M., Grove, M. & Band, L. E. Evaluation of the national land cover database for hydrologic applications in urban and suburban Baltimore, Maryland 1. JAWRA Journal of the American Water Resources Association 46, 429–442 (2010).
Levin, K. et al. Designing and preparing intended nationally determined contributions (INDCs). World Resources Institute: Washington, DC, USA (2015).
USDA. Inflation Reduction Act, 2023).
Lan, C., Wen-Tao, W. & Liu-Cai, Z. The role and development of Green Climate Fund in the global climate governance system. Advances in Climate Change Research 15, 326 (2019).
Albert, C. et al. Addressing societal challenges through nature-based solutions: How can landscape planning and governance research contribute? Landsc Urban Plan 182, 12–21 (2019).
Locke, D. H. et al. Residential housing segregation and urban tree canopy in 37 US Cities. npj urban sustainability 1, 1–9 (2021).
Iungman, T. et al. Cooling cities through urban green infrastructure: a health impact assessment of European cities. The Lancet 401, 577–589, doi:10.1016/S0140-6736(22)02585-5 (2023).
Hartig, T., Mitchell, R., De Vries, S. & Frumkin, H. Nature and health. Annual review of public health 35, 207–228 (2014).
Romitti, Y., Sue Wing, I., Spangler, K. R. & Wellenius, G. A. Inequality in the availability of residential air conditioning across 115 US metropolitan areas. PNAS Nexus 1, pgac210 (2022).
US Census Bureau. 2010 Census Urban and Rural Classification and Urban Area Criteria. (US Census Bureau, Washington, DC. Online at https://www.census.gov/programs-surveys/geography/guidance/geo-areas/urban-rural/2010-urban-rural.html, 2018).
EPA. Greenhouse Gas Equivalencies Calculator, 2022).
McDonald, R. I. Ecosystem service demand and supply along the urban-to-rural gradient. Journal of Conservation Planning 5, 1–14 (2009).
McDonald, R. I. et al. Funding trees for health: An analysis of finance and policy actions to enable tree planting for public health. (The Nature Conservancy, Washington, DC, 2017).
McDonald, R. I., Biswas, T. & Spotswood, E. Throwing Shade at the Green New Deal. LA + Interdisciplinary Journal of Landscape Architecture 15 (2022).
Chakraborty, T., Hsu, A., Manya, D. & Sheriff, G. A spatially explicit surface urban heat island database for the United States: Characterization, uncertainties, and possible applications. ISPRS Journal of Photogrammetry and Remote Sensing 168, 74–88 (2020).
Nowak, D. J. & Greenfield, E. J. Declining urban and community tree cover in the United States. Urban Forestry & Urban Greening 32, 32–55 (2018).
Manson, S., Schroeder, J., Van Riper, D. & Ruggles, S. Ipums National Historical Geographic Information System: Version 17.0., (IPUMS, Minneapolis, MN, 2022).
Dewitz, J. (ed U.S. Geological Survey) (https://doi.org/10.5066/P9KZCM54, 2021).
Gorelick, N. et al. Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote Sens Environ 202, 18–27 (2017).
Nowak, D. J. & Greenfield, E. J. US urban forest statistics, values, and projections. J For 116, 164–177 (2018).
McPherson, E. G., van Doorn, N. S. & Peper, P. J. Urban tree database and allometric equations. Gen. Tech. Rep. PSW-GTR-253. Albany, CA: US Department of Agriculture, Forest Service, Pacific Southwest Research Station. 86 p. 253 (2016).
Nowak, D. J., Greenfield, E. J., Hoehn, R. & Lapoint, E. Carbon storage and sequestration by trees in urban and community areas of the United States. (USDA Forest Service, Washington, DC, 2013).
Zhang, P., Bounoua, L., Imhoff, M., Wolfe, R. & Thome, K. Comparison of MODIS land surface temperature and air temperature over the continental USA meteorological stations. Canadian Journal of Remote Sensing 40, 110–122 (2014).
NOAA. (ed National Oceanic and Atmospheric Administration) (Online at: https://www.ncdc.noaa.gov/data-access/land-based-station-data/land-based-datasets/global-historical-climatology-network-ghcn, 2017).
Venter, Z. S., Chakraborty, T. & Lee, X. Crowdsourced air temperatures contrast satellite measures of the urban heat island and its mechanisms. Science Advances 7, eabb9569 (2021).
Zhao, J., Zhao, X., Wu, D., Meili, N. & Fatichi, S. Satellite-based evidence highlights a considerable increase of urban tree cooling benefits from 2000 to 2015. Glob Change Biol n/a, doi:https://doi.org/10.1111/gcb.16667.
Bobb, J. F., Peng, R. D., Bell, M. L. & Dominici, F. Heat-related mortality and adaptation to heat in the United States. Environ Health Perspect 122, 811 (2014).
Gronlund, C. J., Zanobetti, A., Schwartz, J. D., Wellenius, G. A. & O’Neill, M. S. Heat, heat waves, and hospital admissions among the elderly in the United States, 1992–2006. Environ Health Perspect 122, 1187 (2014).
Barreca, A., Clay, K., Deschenes, O., Greenstone, M. & Shapiro, J. S. Adapting to climate change: The remarkable decline in the US temperature-mortality relationship over the twentieth century. Journal of Political Economy 124, 105–159 (2016).
Murphy, K. M. & Topel, R. H. The value of health and longevity. Journal of political Economy 114, 871–904 (2006).
Kniesner, T. J., Viscusi, W. K., Woock, C. & Ziliak, J. P. The value of a statistical life: Evidence from panel data. Review of Economics and Statistics 94, 74–87 (2012).
Zivin, J. G. & Shrader, J. Temperature extremes, health, and human capital. The Future of Children 26, 31–50 (2016).
Knowlton, K., Rotkin-Ellman, M., Geballe, L., Max, W. & Solomon, G. M. Six climate change–related events in the United States accounted for about $14 billion in lost lives and health costs. Health Affairs 30, 2167–2176 (2011).
Schmeltz, M. T., Petkova, E. P. & Gamble, J. L. Economic Burden of Hospitalizations for Heat-Related Illnesses in the United States, 2001–2010. International journal of environmental research and public health 13, 894 (2016).
Ray, K. N., Chari, A. V., Engberg, J., Bertolet, M. & Mehrotra, A. Opportunity costs of ambulatory medical care in the United States. The American journal of managed care 21, 567–574 (2015).
EPA. Report on the Social Cost of Greenhouse Gases: Estimates Incorporating Recent Scientific Advances. (US Environmental Protection Agency, 2022).
Rennert, K. et al. Comprehensive evidence implies a higher social cost of CO2. Nature 610, 687–692 (2022).
US Bureau of Labor Statistics. (ed US Bureau of Labor Statistics) (Washington, DC., 2022).

Supplementary Table S1 is not available with this version

(Not answered)

Current inequality and future potential of US urban tree canopy cover for reducing heat-related mortality, morbidity and electricity consumption

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Discussion

Methods

Declarations

References

Supplementary Table

Additional Declarations

Status:

Journal Publication

Version 1