The week that Death Valley National Park rangers realized David Kelleher was missing, temperatures soared to 123 degrees Fahrenheit. On Wednesday, June 8, a ranger spotted the 67-year-old’s car in the Zabriskie Point parking lot, where sweeping views of the badlands and salt flats make it one of the most popular stops in the park, which straddles the California-Nevada border. On Saturday night, after a 122-degree day, the same ranger noticed that Kelleher’s car was still there—with a note inside reading “out of gas”—and launched an investigation.
sHeat that weekend slowed the search. In temperatures over 115 degrees, helicopters can’t generate the lift to fly safely, and daytime ground searches are dangerous, too. When they could, searchers focused on the trail system nearby. But he wasn’t found until early the following week—park visitors discovered him about 2.5 miles away from his car, his body hidden by a tree just 30 feet off the main road.
When Abby Wines, the park’s public information officer (PIO), prepared the press release the next day, she wrote that it seemed he’d been walking toward Furnace Creek, where there’s an inn, a visitor center, and a gas station. She also noted how hot it had been, figuring readers would make the connection with that week’s heat wave. But, as in most park deaths she thinks are likely heat related, she couldn’t formally say that it was. This is because it is tough, sometimes impossible, to identify as a cause of death with an autopsy. Investigations of fatalities like these by medical examiners can take weeks and even months, if they happen at all, leaving the public no better informed about just how dangerous local conditions are and who is at risk of illness or death on the trail.
It’s part of a problem that plagues heat-fatality data nationwide. These deaths occur in parks, cities, stadiums, farm fields, senior-living apartments, and homeless camps, both during extreme heat events but also on regular hot days. Most who die from heat-related illness are never officially counted as having died from hot conditions, which epidemiologists, coroners, local and state governments, the Environmental Protection Agency, and the Centers for Disease Control and Prevention all say hampers the public’s understanding of the dangers of heat and stymies efforts to prevent further deaths.
Kelleher’s was the second such death in a matter of weeks, following John McCarry, 69, whose body was found June 1. In 2021, of six hiker fatalities in Death Valley, park press releases noted the high temperatures at the time related to three. A fourth, a 27-year-old woman who died hiking in November, was speculated by some media as a heat-related fatality, but her cause of death was never publicized.
“It becomes very awkward for me as a PIO,” Wines says. “There’s no sign of foul play. That’s the only thing I can say. There’s no sign of suicide or someone attacking them. But the difference between a heat-related medical event and a medical event that’s just your time to go= and you happen to be in the middle of nowhere—usually that nuance isn’t very clear.”
Outdoor adventurers of all stripes are used to databases, logs, and publications that help them (and researchers) understand how people die outside so they might avoid a similar fate, including the national avalanche-fatality database, the American Alpine Club’s annual report Accidents in North American Mountaineering, the American Whitewater Accident Database, and more.
Heat, the “leading weather-related killer in the United States,” according to the EPA, has no such database. But as public-health experts around the country call for a better accounting of heat-related illnesses and deaths, efforts to do so are taking hold in places from Phoenix to California to the CDC’s offices in Atlanta. These initiatives could help us better understand a phenomenon that doctors and climate scientists agree is getting deadlier, preventing future deaths from occurring.
In the 1990s, two of the nation’s worst heat waves swept the country, killing 118 people in Philadelphia in ’93 and 739 in Chicago in ’95. In reports on both disasters, the CDC noted that its understanding of the deaths was limited, because medical examiners had “no uniform definition for heat-related death.”
CDC officials later estimated the number of heat-related deaths by comparing those numbers to average deaths from previous years—a practice called “excess death analysis,” which is still the preferred method to estimate fatalities during disasters, including the COVID-19 pandemic. But the incomplete data meant that officials in Chicago couldn’t specify which deaths were heat related or whether certain groups had been impacted disproportionately and why. It also meant that, outside of large disasters like these, officials had no reliable understanding of how many people die from heat, because it’s so hard to definitively identify with an autopsy, and because not everyone who dies of heatstroke gets one in the first place.
In response, in 1997, the National Association of Medical Examiners (NAME) published “Criteria for the Diagnosis of Heat-Related Deaths” in The American Journal of Forensic Medicine and Pathology. But even these criteria have limitations. If a person’s core temperature before death is at least 105, the criteria states, “cause of death should be certified as heat stroke or hyperthermia.” Lower readings can still qualify if witnesses or blood tests before death verify symptoms of heatstroke, such as confusion, low sodium, or heat-damaged muscle proteins leached into the blood.
If a person is found dead, which is often the case on trails and during heat waves, the criteria are squishier. An autopsy can’t identify those same conditions. Environmental factors might be unknown. In older individuals with underlying conditions, which are often exacerbated by heat, diagnosis can be even trickier.
Kathryn Pinneri, current president of the NAME and director of forensic services for Montgomery County, Texas, says that, given the hypothetical of a hiker who died in Death Valley when conditions were known to be hot but the body wasn’t found for several weeks, she would probably classify their death as “undetermined.”
“Unless there’s something really clear, like a specific note or text message saying, ‘I forgot my water and it’s 120 degrees,’ most of us would classify those deaths as undetermined if there was no definable cause of death,” she says.
The CDC, which pulls from coded death-certificate data, reports that about 700 Americans die each year from heat-related illnesses. A highly regarded excess death analysis published in 2020 by Environmental Epidemiology puts that number at more than 5,600.
Greg Wellenius, an environmental epidemiologist and an author of that analysis, says that even those numbers are “probably an underestimate,” pointing to another research group that calculates it closer to 10,000.
“That number—whether it’s 6,000, 10,000, or 15,000—is in the ballpark of something we should pay attention to,” he says. “It’s worth investing in better understanding: Who is at risk, when are they at risk, and, importantly, what can we do about it?”
The nature and scale of heat-related deaths makes them difficult to compare with other environmental accidents, but the way we handle avalanche deaths stands out as a case study in what’s possible when the resulting deaths are both well documented and widely communicated to the public.
At their most basic, heat and avalanches are common natural phenomena that regularly cause preventable deaths. Researchers who study both talk similarly about the value of learning from previous incidents, but while heat has remained largely underestimated by the public and the data, the systems in place to forecast avalanches, educate the public, and document exactly who dies in such events and why is a public-health success story.
The national avalanche-fatality database, maintained by the Colorado Avalanche Information Center (CAIC) in partnership with the U.S. Forest Service, grew out of an effort to understand a new and pressing problem: the exploding popularity of skiing in the mid-20th century in the Mountain West, which placed more recreationalists at risk. The Forest Service developed avalanche-safety and forecasting programs to protect skiers—along with mountain-town residents, drivers, and workers—and a group of forecasters “saw the value in starting to better document” incidents, says Karl Birkeland, director of the Forest Service’s National Avalanche Center.
The first Snowy Torrents, which publishes detailed accounts of every fatal avalanche accident on record, came out in 1968 and has since grown into a database that records accidents from the mid-1800s through the present day.
“It resonates with people when you tell them exactly what happened in an accident, so hopefully they can avoid it,” says Ethan Greene, who runs the CAIC. “People relate much better to real-world narratives than a kind of esoteric description of a potential threat.”
Today, he says, reports are based on data collected as quickly as possible by staff of the nearest avalanche center, who travel to document snow and weather conditions, narratives from witnesses and survivors, and any available information from first responders and coroner reports. Although resource intensive, this kind of information has proven invaluable in the effort to increase public understanding of conditions and risks during the winter season and has helped researchers, policymakers, and rescuers alike understand long-term trends about how and why people die in avalanches.
Birkeland says that since the mid-1990s, when their reports and forecasts became easily accessible online, annual deaths have stabilized at around 25 to 30 per year despite an estimated 12-fold increase in recreation. “We can use that as a rough measure of the effectiveness of our avalanche forecasting and education programs,” he says. Education, outreach, and data collection about heat deaths could help equip people to better navigate that public-health crisis, too.
Public understanding of the deadly nature of extreme heat seems to be improving. In the aftermath of the 2021 Pacific Northwest heat wave, The New York Times published an excess death analysis showing that 600 people had died, while death-certificate data recorded less than 200. An investigation in the Los Angeles Times later that year noted that heat deaths in California are as much as six times higher than official counts.
Now, Wellenius says, “there’s a huge movement to address the problem.”
The epidemiologist points to the city of Phoenix, which last year created an Office of Heat Response and Mitigation, and to Miami, which appointed the world’s first chief heat officer. This spring, California governor Gavin Newsom’s office launched an extreme-heat action plan; one goal is to “modernize the Electronic Death Registration System to register heat-related deaths” in order to help prevent them. And the CDC says it continues to improve its new Heat and Health Tracker, a dashboard launched in 2021 that tracks weather conditions and heat-related ER visits.
David Hondula, who leads the new heat office in Phoenix, says that his work advocating for solutions and preventing deaths is possible because the Maricopa County Public Health Department has a complete grasp of who is getting sick and dying from heat. Because that department monitors hospital discharge data for heat-related diagnoses using an especially broad set of criteria, “we’re able to precisely understand” who is dying, how, and where, he says. He knows how many of the 339 who died in 2021 were houseless, had underlying conditions, or were out hiking on Camelback Mountain, a popular area hike where average temperatures hover above 100 degrees four months of the year.
This means he can direct hydration and cooling stations where data shows they’re most needed, advocate for affordable housing, because more people die without it, and ensure appropriate signage on trails where conditions and rescues are particularly dangerous in heat.
Back in her office in Death Valley National Park, Wines has nothing like this. In fact, she says, she hardly has any data at all. She recalls a recent call from another reporter, who requested annual stats on heat-related emergencies and deaths.
“It looks like we’re just avoiding the question, but we don’t have everything coded that way in our system,” Wines acknowledges. “I would love it if it was a clear thing and we could say: this many people on average die because of heat, and we have this many medical responses by law enforcement, and it costs the taxpayers this much money.” But she doesn’t have that data.
If she did, she says, “I think it would be a lot easier to tell people how dangerous it is.”