Heat kills more Americans each year than floods, lightning, tornadoes, and hurricanes combined. Federal mortality data show the toll reaches into the hundreds annually, yet the victims rarely make national headlines. Unlike a tornado that flattens a neighborhood or a hurricane that triggers mass evacuation, extreme heat builds silently, and the people most at risk often have no warning that their bodies are failing until it is too late.
Why silent heat deaths outpace every other weather threat
The National Weather Service states it plainly on its heat safety page: “Heat is the number one weather-related killer in the United States.” That ranking holds year after year in federal storm and mortality records, and it persists because heat does not produce the visible destruction that drives emergency response and media coverage. A tornado leaves debris. A flood leaves waterlines. Heat leaves a person who stopped sweating, lost consciousness, and never called for help.
A CDC analysis of death certificates from 2004 through 2018 found that heat-related deaths averaged 702 per year during that period. Of those, 415 listed heat as the underlying cause and 287 listed it as a contributing cause, meaning a pre-existing condition such as heart disease or kidney failure became fatal only because the body could not cool itself. That distinction matters because “heat-contributed” deaths are the ones most likely to go unrecognized by families and even by first responders. A person with congestive heart failure who dies during a heat wave may never be counted in a storm-damage tally, but the heat is what tipped the scales.
Separate CDC surveillance of the National Vital Statistics System covering 1999 through 2020 recorded deaths involving exposure to excessive heat by sex and showed that in some years the count exceeded 1,000 people per year, depending on how broadly contributing causes were counted. The gap between the lower and higher estimates reflects a measurement problem that runs through the entire field: different agencies, different definitions, and different thresholds for calling a death “heat-related.”
Federal data and local autopsies tell different stories
Two primary federal systems track heat fatalities, and neither captures the full picture on its own. NOAA’s Storm Events Database records weather events including heat, drawing on National Weather Service office reports and local emergency management. It is the backbone of the comparative claim that heat outpaces all other weather hazards. But it depends on events being reported and classified in real time, which means slow-onset heat deaths, especially those that occur indoors over several days, can slip through if no one connects them to a specific heat episode.
The CDC’s Multiple Cause of Death portal works from death certificates filed months or years after the fact, using ICD-10 codes to flag excessive heat exposure. It captures a wider net of cases but cannot tell you whether the person had air conditioning, whether they were unhoused, or whether a cooling center existed nearby. Those details live only in local medical examiner files, ambulance run sheets, and police reports.
Maricopa County, Arizona, stands out because its Department of Public Health publishes unusually detailed annual heat reports that classify each death as heat-caused or heat-contributed and record risk factors such as homelessness and lack of air conditioning. The county’s 2025 heat report draws on preliminary medical examiner investigative findings alongside death certificates, producing a case-level picture that no national dataset replicates. In Maricopa County, the pattern is consistent: many victims are found indoors, alone, with broken or absent cooling equipment, and with chronic health conditions that made them especially vulnerable.
This local detail is exactly what federal datasets lack. The CDC WONDER system covers multiple causes of death from 1999 through 2020 but contains no field for housing status, indoor temperature, or access to cooling. The Storm Events Database narratives do not include individual-level data on medication use, mental health, or whether the person had previously been warned about heat risk. Without those variables, researchers cannot answer the question that matters most for prevention: did a specific intervention, such as opening a cooling center or distributing portable air conditioners, actually reduce deaths among the people at highest risk?
Cooling centers and the unanswered prevention question
A reasonable hypothesis holds that counties adding temporary cooling centers after 2020 should have seen a measurable drop in heat-contributed deaths among unhoused residents compared with similar counties that did not add centers. Testing that claim requires three things: county-level mortality data broken out by housing status, records of when and where cooling centers opened, and temperature data to control for whether one county simply had a milder summer. No single public dataset currently provides all three, and even where pieces exist, they are rarely standardized across jurisdictions.
Maricopa County’s heat surveillance program comes closest. Its reports track homelessness as a risk factor and document year-over-year death counts alongside local cooling interventions. Public health officials can see, for example, whether a spike in outdoor deaths coincides with a stretch of nights that remained dangerously hot, or whether indoor deaths cluster in neighborhoods with older housing stock and high utility shutoff rates. But Maricopa is one county in one state, and its desert climate and demographics do not generalize easily to Philadelphia, Chicago, or Houston, where humidity, housing, and social services look very different.
A recent federal climate-health assessment synthesized evidence that climate change is expanding the number of dangerously hot days and raising exposure for the same vulnerable groups, including older adults, outdoor workers, and people without reliable housing. Yet that assessment supplies no post-2020 county-level mortality counts in a format that would allow the kind of matched comparison needed to evaluate specific policies. Researchers can say with confidence that extreme heat is becoming more frequent and more intense, and that deaths are concentrated among socially and medically vulnerable residents. They cannot yet say which mix of interventions – cooling centers, tree planting, utility shutoff moratoria, targeted wellness checks – delivers the greatest reduction in deaths per dollar spent.
That evidence gap leaves local officials making high-stakes decisions with incomplete information. Cities open cooling centers but struggle to reach people who distrust government buildings or lack transportation. Nonprofits distribute fans and air conditioners, but some recipients cannot afford the resulting electricity bills and never plug them in. Utility regulators debate summer shutoff bans without clear numbers on how many heat deaths are linked to power loss. Each measure may save lives, but without integrated data tying individual deaths to specific exposures and services, the impact remains largely inferred rather than measured.
Closing that gap would require aligning federal and local systems around a shared goal: tracking not just how many people die from heat, but how and why. Death certificates could include standardized fields for housing status and indoor environment during the fatal event. Local medical examiners could feed de-identified case summaries into a national registry that links to weather observations and neighborhood-level infrastructure data. Public health departments could then evaluate whether deaths fall in areas served by new cooling centers or remain stubbornly high among people sleeping outdoors or in substandard housing.
Until then, the country’s deadliest weather hazard will continue to be the least visible. The numbers in federal databases confirm that heat quietly kills more Americans than flashier storms, but the stories behind those numbers remain largely hidden in local files. Without bringing those stories into focus, policymakers and communities are left to fight a growing threat with blunt tools and educated guesses, while the most vulnerable residents bear the brunt of a danger they often cannot see or escape.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
