Heat disproportionately kills young people: Evidence from wet-bulb temperature in Mexico
with Danny Bressler, Tereza Cavazos, Radley Horton, Casey Ivanovich, Patrick Kinney, Colin Raymond, Jeff Shrader, Adam Sobel, and Cascade Tuholske
Recent studies project that temperature-related mortality will be the largest source of damage from climate change, with particular concern for the elderly whom it is believed bear the largest heat-related mortality risk. We study heat and mortality in Mexico, a country that exhibits a unique combination of universal mortality microdata and among the most extreme levels of humid heat. Combining detailed measurements of wet-bulb temperature with age-specific mortality data, we find that younger people who are particularly vulnerable to heat: People under 35 years old account for 75% of recent heat-related deaths and 87% of heat-related lost life years, while those 50 and older account for 96% of cold-related deaths and 80% of cold-related lost life years. We develop high-resolution projections of humid heat and associated mortality and find that under the end-of-century SSP 3–7.0 emissions scenario, temperature-related deaths shift from older to younger people. Deaths among under-35-year-olds increase 32% while decreasing by 33% among other age groups.
Published in Science Advances. Manuscript here.
Health damage from transboundary and domestic air pollution in mixture
with Jaecheol Lee and Solomon Hsiang
Particulate matter (PM) is the most clinically important air pollutant. Current studies assume that units of PM originating in different jurisdictions cause the same harm, despite widespread understanding that differing chemical and physical features of PM could generate distinct health effects. Here, we combine an atmospheric model, universal health records, and econometric analysis to provide the first direct evidence that the health impact of PM depends on its origin. We simultaneously measure harm from seven categories of PM within a single population at the nexus of the world's most contentious transboundary air pollution dispute. Because impacts differ by origin, we compute that transboundary sources contribute only 43% of anthropogenic PM load to our study population, but generate >70% of associated respiratory health costs. Our results indicate that PM should be considered a mixture of pollutants of distinct origins, each with a unique measurable impact on human health.
Recent draft here.
Want to talk to the paper instead? Try TransboundaryGPT.
Weather and the safety of U.S. railways
with Xinming Du
Railway safety is affected by the weather. We quantify these effects by leveraging a comprehensive dataset on railway safety incidents in the United States spanning 1997–2019. Though weather conditions are noted as a primary or contributing factor to 2.2% of railway safety incidents during our data period, we find that weather causes closer to 8.5% of all rail safety incidents—four times the documented amount. Both heat and cold cause elevated incident counts, with effects especially strong for incidents leading to injuries or deaths. Exposure to a daily average temperature over 30ºC (86ºF) leads to a 9.5% increase in the number of rail safety incidents, a 27% increase in the number of incidents leading to a casualty, and 25% and 57% increases injuries and deaths—effects net of any operational adjustments made to mitigate these effects. Extreme cold and extreme precipitation also affect safety. We find that locations are adapted to their local climate, with, for example, warmer places exhibiting a weaker relationship between heat and incident count. Further, past exposure to hazardous weather leads to fewer accidents future accidents, perhaps demonstrating learning. The numbers of injuries and deaths associated with rail system weather exposure may suggest a role for enhanced rail safety regulations and adaptation funding to protect critical infrastructure.
See a recent draft of our paper here.
Implications of wildfire-derived PM2.5 for U.S. air quality regulation
with Marissa Childs, Mariana Martins, Sam Heft-Neal, Minghao Qiu, and Marshall Burke
Growing wildfire activity across North America generates large amounts of smoke that can travel long distances. Characterizing the influence of this smoke on surface air quality is critical for regulation of air quality and protection of public health. Here we provide granular, daily estimates of smoke fine particulate matter (PM2.5) concentrations across the contiguous U.S. and use them to assess the influence of wildfire smoke on surface PM2.5 from 2006 to 2023, using a combination of surface measurements, satellites, and machine learning. Each year from 2020 to 2023, population-average smoke PM2.5 exposures were 2.6–6.7 times higher than the 2006 to 2019 average and exposure periods were twice as long. Despite wildfire smoke being historically more common in the Western U.S., the worst 5 days for population-average smoke exposure in our sample period all occurred in 2023, a year with limited Western exposure. We estimate that wildfire smoke is pushing 34% of monitoring stations above the recently-updated ambient air quality standards, necessitating increased use of extreme event exemptions to remain within regulatory limits; we show that such use is already rare on attainment-relevant days and could become increasingly challenging. Absent wildfire smoke, we estimate that PM2.5 concentrations would have continued to improve throughout the contiguous U.S.
See a preprint here. Access a prepublication version of the data here.
Intensifying risk of mass human heat mortality if historical weather patterns recur
with Chris Callahan, Jared Trok, Carlos Gould, Sam Heft-Neal, Noah Diffenbaugh, and Marshall Burke
The potential death toll of worst-case extreme heat events is crucial for climate risk analysis and adaptation planning. We estimate this quantity for Europe using machine learning to calculate the intensity of historical heat waves if they occur at present or future global temperatures, combined with empirical exposure-response functions to quantify the resulting mortality. Each event is projected to generate tens of thousands of excess deaths. For example, if July 1994 or August 2003 meteorological conditions recur at the current global temperature anomaly of 1.5°C, we project 14,000 or 17,300 excess deaths across Europe in a single week, respectively. At 3°C, mortality rises to 26,800 or 31,500 per week. These death rates are comparable to peak COVID-19 mortality in Europe and are not substantially reduced by ongoing climate adaptation. Our results suggest that avoiding mass heat mortality in Europe will require significant and novel adaptation to heat.
See a preprint here.
Air pollution, weather, and criminal courts
Environmental conditions, such as weather, affect emotion and cognition. These effects have wide-ranging consequences for social and economic systems. Some evidence suggests effects may also extend to outcomes of the criminal justice system. Here, I investigate the effect of air pollution and ambient temperature on verdicts and sentence severity for criminal cases in the United States. I find that criminal cases processed during hot days have significantly higher odds of ending in a guilty verdict. For example, an average criminal trial with daily temperatures of 20ºC leads to a guilty verdict about 66% of the time, while daily temperatures of 25ºC are plausibly associated with a guilty verdict around 89% of the time. However, conditional on verdict, temperatures do not appear to significantly affect sentencing severity. Conversely, I find strong evidence that lower air quality causes judges to issue more severe sentences conditional on guilt and that levels of most criteria air pollutants are associated with important changes in both verdicts and sentencing. Across the sample period (May 1998 to May 2003), solely the portion of air pollution above the EPA's limit for “healthy air” led to the annual issuance of an additional $16 million (2021 dollars) in criminal fines and the effective removal of an additional 49,000 people from society via incarceration. While these effects amount to a significant potential cost of extreme heat and air pollution, their policy implications are unclear.
See a recent draft of this paper here.
The effect of landfilled waste on infant health
Human activity in the United States alone generates around 600 billion pounds of municipal solid waste annually, an astounding figure that has grown more than threefold since 1960. By the end of this decade, humanity may generate upwards of 10 trillion pounds of waste annually. The vast majority of these wastes are placed in sanitary landfills, engineered structures that hold compacted waste and manage its effluent (“leachate”). While these structures are designed to prevent leaks into the surrounding environment, precisely monitoring these large structures is impractical. Compounding this uncertainty, the rates and stages of leachate production within landfills are poorly understood and depend on a complex set of factors, including local climatic conditions as well as a landfill’s content, size, and age. In this study, I combine satellite imagery and machine learning to construct a panel of the quantities of waste at most large landfill sites throughout the United States. I then construct a national dataset of the locations of public water system intakes, which I use to generate a measure of landfilled waste exposure intensity. By comparing outcomes for individuals whose water is drawn upstream and downstream of landfill sites experiencing precipitation shocks, I determine that a one standard deviation increase in upstream landfilled waste is associated with higher infant mortality and increases in the portion of births that are low weight. I discuss implications for waste management policy and the assignment of liability for landfill malfunction through the U.S. Resource Conservation and Recovery Act.
Early stage project—reach out!
Ocean acidification and global fish populations
By 2100 under the SSP 3-7.0 GHG emission scenario, the acidity of the Earth's oceans are projected to roughly double. The social costs of this process are unquantified. Two economically important categories of organisms may experience large impacts: plankton, krill, and mollusks, which may be unable to as readily form carbonate exoskeletons, leading to slowed growth and higher mortality; and finfish, whose sense of smell is affected by ocean pH. In this study, I combine a remotely sensed measure of local ocean acidity, temperature, and deoxygenation; species-specific life histories for most economically important finfish; and a measure of fish catch per unit effort to estimate the effect of anthropogenic carbon emissions on fishery productivity, net of effort-based adaptation. Ocean acidity is often correlated with local weather, so I leverage a change in international marine shipping regulations that led to the rapid and widespread adoption of sulfur dioxide scrubbers on long-distance marine shipping vessels. Open-loop scrubbers percolate emission plumes through seawater, which becomes highly acidic before being released, thus implementation of this regulation created a exogenous shock to ocean acidity. Using this shock, I determine that perhaps 20% of the decline in fish stocks since the 1970s may have been due to the acidification of the oceans alone.
Early stage project—reach out!
For information on my other research, see my CV.