Was It Really 134 Degrees? The Doubtful Story of Earth’s Hottest Day

Outside the visitor center at Furnace Creek, in the bottom of Death Valley, there is a big digital thermometer that has become a favorite photo stop. People line up in the afternoon to stand next to it when the numbers climb into the 120s, the way other travelers pose beside a mountain summit marker. On July 12, 2020, two masked park rangers stood on either side of it while it read 129°F. The sign’s lower line said 59°C, which is wrong: 129°F is about 54°C, and the Park Service, publishing a photo of the same sign the next month, added a dry note that “the Celsius reading should display 54 degrees.” It is a small glitch, but it is a fitting one for this valley, because the most famous number ever written down there may be wrong by a great deal more.

That number is 134°F, or 56.7°C, recorded at a place called Greenland Ranch on July 10, 1913. It is still the official highest air temperature ever measured on Earth, listed by the World Meteorological Organization and celebrated on park signs, T-shirts and souvenirs. It has also been doubted for most of the century since it was written down. In January 2026, three atmospheric scientists published the most detailed case yet against it in the Bulletin of the American Meteorological Society. Their verdict: the real high that day was probably about 120°F, give or take two degrees, and the world record should be rescinded.

Earlier this month we wrote about the sailing stones of the Racetrack Playa, a Death Valley mystery that was solved when researchers finally put cameras and GPS trackers on the rocks and watched. The heat record is the opposite kind of puzzle. Nobody can go back to July 1913 with a better thermometer. What is left is paper: monthly weather forms in an observer’s handwriting, letters between a borax company’s managers, a few old photographs, and the readings from every other weather station in the desert. Read carefully, that paper tells a story about an oasis, a ranch foreman, a valley with a reputation to live up to, and a number that may have been too good to check.

Why Death Valley is so hot in the first place

None of the doubt is about whether Death Valley is hot. It is, by any honest measure, the hottest place in North America and one of the hottest on the planet. The valley floor reaches 282 feet below sea level at Badwater Basin, the lowest point on the continent. It is long and narrow, walled in by steep mountains, and it sits behind four mountain ranges that wring the moisture out of Pacific storms before they arrive. What little air reaches the valley floor has already sunk a long way, and sinking air warms as it is compressed. The sun heats the bare rock and salt; the rock heats the air; the walls keep the hot air from mixing away.

The park’s own record book reads like a list of dares. On July 15, 1972, a ground temperature of 201°F was measured, on a day when the air was 128°F. In 2001 the valley went 154 consecutive days with highs of 100°F or more. In 1917 it had 43 straight days at 120°F or above. There was no measurable rain at all in 1929 or in 1953. On the night of July 12, 2012, the temperature never fell below 107°F, the warmest low ever recorded there, and that day’s 24-hour average of about 117.5°F is still cited as the hottest on record anywhere.

The landscape matches the numbers. Salt pans, badlands, gravel fans and dunes, including the Eureka Dunes, which hum when the sand slides, fill the basin. From orbit the valley looks like a long scar between dark mountains; Landsat 1, the first satellite in the long-running Landsat series, photographed it in false color in 1976 with Furnace Creek near the center.

All of that is real. The question is narrower, and in a way more interesting: could the air in a properly sheltered thermometer at Greenland Ranch really have reached 134°F on one particular Thursday in July 1913, when nothing like it has been measured there since?

A valley that grew its own legends

Death Valley got its name from fear before it had any thermometers. Late in 1849, a party of gold seekers trying a shortcut to California became stranded there. Two young men, William Lewis Manly and John Rogers, walked nearly 300 miles to Mission San Fernando for supplies and came back 26 days later to lead the survivors out. As the last of them left, someone reportedly called out, “Goodbye, Death Valley,” and the name stuck.

The legends grew fastest after the borax miners arrived. In 1883 the Harmony Borax Works began refining borax scraped from the valley floor near Furnace Creek, and it became famous for the twenty-mule teams, really eighteen mules and two horses, that hauled wagons of borax 165 miles across the desert to the railroad at Mojave on trips that took about ten days. The same company set up a ranch at Furnace Creek that year and planted it with alfalfa. They called it Greenland Ranch. It was, quite literally, an oasis: a square of irrigated green on a pale desert fan.

The stories that came out of the valley in those years were spectacular. A mineral surveyor named James McGillivray told a writer in the early 1890s that his Signal Service thermometer, hung in the shade on the side of a shed, “repeatedly registered 130,” and that for forty-eight hours in 1883 it never once fell below 104. A local chronology credits James Dayton, the ranch caretaker, with a reading of 136°F on July 3, 1890. An 1892 book on the borax deserts reported that a thermometer “hanging under the wide veranda on the north side of the adobe house on the Death Valley ranch, has registered 137 degrees.” Scientific American, in 1885, had already called the valley the Puit d’enfer, the pit of hell, and lamented that no one could survive there long enough to keep continuous observations.

When trained observers actually measured the heat, the numbers came out lower. The Wheeler survey recorded 121°F in 1875. In 1891 the chief of the newly organized U.S. Weather Bureau, Mark Harrington, set up a temporary first-order station near the borax works, about two miles northwest of the mouth of Furnace Creek, and ran it from May through September. An observer named John H. Clery took hourly readings with a thermometer in a proper shelter. The highest temperature of the whole summer was 122°F, reached on three days in a row from June 30 to July 2 and again in late August. It was miserable heat; Clery’s assistant, R. H. Williams, “succumbed to the heat soon after his arrival” and had to be taken out for medical care. But it was not 136.

The historian Richard Lingenfelter later gave his history of the region the subtitle A Land of Illusion, and that idea is the thread the new study pulls on. By the time the Weather Bureau sent a permanent set of instruments to Greenland Ranch, the people living there already knew, or believed they knew, how hot it got. The official equipment was about to disappoint them.

An oasis with a weather station

The Weather Bureau’s cooperative station at Greenland Ranch opened in 1911. It was a standard arrangement for the time. A white wooden instrument shelter with louvered sides, the kind that protects thermometers from direct sun and reflected heat while letting air flow through, was set up by the ranch superintendent, Fred Corkill, who had kept weather records before in Nevada. Inside were a maximum thermometer, a minimum thermometer and other instruments, read each day by whoever was working the ranch. Nobody from the Weather Bureau appears to have visited the station to inspect it or train the observers until March 1924, thirteen years after it was installed.

Where the shelter stood mattered enormously, and on that point the Weather Bureau’s own forecaster in San Francisco, George Willson, left a remarkably candid description in 1915. “The location for the instruments was carefully selected,” he wrote, “the shelter being placed over an alfalfa sod, the floor about 4 feet above the ground, the shelter door facing north.” Because the desert was so dry, the fields needed heavy irrigation, and Willson thought “the cooling by evaporation from surrounding damp ground and live vegetation is probably sufficient to lower the readings of the instruments several degrees.” In plain terms, the official thermometer was sitting in the coolest spot for miles.

The ranch noticed. In June and July of 1911, Corkill’s correspondence with the Pacific Coast Borax Company expressed disappointment that the official readings were so much lower than everyone expected. The men had grown used to the thermometer on the veranda. That thermometer, as far as anyone can tell, was of unknown make, unknown calibration and unknown exposure, and it hung on a building whose north wall would have caught the late-afternoon summer sun.

The first daily observer was the ranch caretaker, Thomas Osborn. He kept the forms for about a year, until he caught typhoid in 1912; he died later that year. From September 1912, the job passed to the new ranch foreman, Oscar A. Denton, who would keep the Greenland Ranch weather record through eight summers, until early 1921.

Oscar Denton’s July

Denton was not a scientist, and he did not pretend to be. He was a loner who looked hard at strangers. The writer Dane Coolidge, who visited the ranch in the 1910s, wrote in his 1937 book Death Valley Prospectors that Denton “had an unpleasant experience during the Mexican Revolution and he had a special reason for being suspicious of strangers, since there were men who had sworn to kill him.” He had retreated, Coolidge wrote, to the most isolated place in North America. In summer, when the borax works and the ranch largely shut down, he was often the only person there. Corkill later said Denton had trouble using the tables that turn wet- and dry-bulb readings into humidity. The writer Frank Crampton remembered him as “a good listener” who “passed the latest stories and the best lies to the next to come.”

The summer of 1913 was hot across the Southwest, and early July brought a genuine heat wave. But the numbers Denton entered for the second week of the month were unlike anything the station had produced before or would produce for the rest of its existence. From July 7 to July 14 his maximum temperatures read 127, 128, 129, 134, 129, 130, 131 and 127°F. Each of those eight days equals or tops every other maximum in the station’s fifty-year record, except for one questionable 129°F in 1960. The minimums, by contrast, were close to normal, between 85 and 93°F.

There was a problem with that 134. The station’s maximum thermometer was graduated only up to 135°F. When Denton sent in the July form, Willson noted, the observer added that “he doubted if the record was sufficiently high because other ordinary thermometers at the ranch showed a much higher temperature.” In other words, the man who recorded the world’s hottest day believed his official instrument was reading too low.

Willson was puzzled for a different reason. He had studied the daily weather maps, he wrote, looking for something that would explain the extraordinary heat, “but it is doubtful if a sufficient cause was found. The weather type was that which always causes high temperatures over the south Pacific coast district, it was not unusually pronounced, and did not give record temperatures in any other portion of California.” The pattern was a normal hot summer pattern. Only Greenland Ranch had gone off the scale.

The trouble with the paperwork

Over the past decade, the weather historian Christopher Burt and the meteorologist William Reid have gone through Denton’s monthly forms line by line, and they found a pattern of oddities that goes well beyond one week. When rain fell at the ranch, the amounts Denton recorded were almost always 0.01, 0.10, 0.20 or 0.30 inches, as if rounded to whatever was convenient. In September 1914 he entered fifteen consecutive maximums of either 109 or 110°F, a run that no nearby station supports. In March 1913, when a cool, windy storm swept the region and dropped temperatures everywhere by about ten degrees, Greenland Ranch reported warming.

The newest paper adds more. Many of the “set max” readings Denton recorded at the evening observation in 1913 to 1915 are higher than the next day’s maximum, which should be impossible if the thermometer was reset once a day as required. The likely explanation is that he reset it more than once. And in two later Julys the numbers look simply invented: in July 1918, 22 of 31 daily highs were exact multiples of five (105, 110, 115 and so on), and in July 1920, 19 were. Across the 102 Julys from 1923 to 2024, the average is 6.7.

There is also the matter of where the shelter was. Two photographs from around 1920 show the Greenland Ranch shelter standing over bare, natural desert, with no alfalfa field in sight. Photographs from the mid-1920s show it back beside the irrigated field. No station record mentions a move. The researchers suspect Denton shifted the shelter, at least for a while, to get readings more like the ones on the veranda, which would fit his note about the “ordinary thermometers.” They are careful to say that a move to bare ground, by itself, could not produce a 134. They also raise a blunter possibility: that some of the hottest numbers of July 1913 never came from the shelter at all.

None of this proves what happened on July 10, 1913. Denton is long gone, and the researchers themselves call every explanation speculative. What the forms do show is that the station’s early record was kept by one isolated man, with no supervision, at a time when everyone around him believed the official thermometer was too cold.

The statistics of an impossible day

The strongest case against the record does not rest on Denton’s character at all. It rests on how desert heat behaves. On a summer afternoon in a place like Death Valley, the lower atmosphere is churned by convection into a deep, well-mixed layer. Within that layer, temperature falls with height at a fairly steady rate, and it changes slowly through the hottest part of the day; in the 1940s the Army climatologist Arnold Court found that Death Valley shelter temperatures stayed within 2 or 3 degrees of the day’s maximum from about 1:30 to 5:30 p.m. in July. That makes the valley floor unusually predictable. If you know how hot it was at the stations around it and how much higher they sit, you can estimate how hot it was at the bottom.

Court himself was skeptical as early as 1949. Using the first 37 years of annual maximums at Greenland Ranch, he calculated that a 134°F reading should come along about once every 650 years. Burt’s later tally makes the same point. Apart from 1913 and that suspect 1960 reading, the hottest day of each year at Greenland Ranch from 1911 to 1960 ranged from 120 to 127°F, and at the modern Death Valley station from 1961 to 2016 from 119 to 129°F. Across 105 years the average annual maximum was 124.3°F with a standard deviation of 2.2 degrees, which puts 134°F about four and a half standard deviations out. Comparing Greenland Ranch to its neighbors during the July 1913 heat wave, Burt found the surrounding stations would support a peak of about 123 to 126°F, not 134.

The new paper, by Roy Spencer and John Christy of the University of Alabama in Huntsville and William Reid, does this comparison more systematically. The authors took 102 years of July data, from 1923 through 2024, for the higher-elevation stations within about 250 kilometers east of the Sierra Nevada, and measured how well they predict Death Valley’s daily highs. The relationship is tight. Temperature in that region falls by about 4.8°F for every thousand feet of elevation, within a range of 4.2 to 5.4, and the predicted Death Valley highs normally match the observed ones with a day-to-day scatter of about two degrees.

Then they ran the same method backward into 1913. For July 10, the surrounding stations point to a maximum of about 120°F, with a likely range of 118 to 122. That makes the official 134 roughly 12 to 16 degrees too warm. July 13’s reported 131 was too warm by a similar amount, and across July 2 to 18, 1913, Greenland Ranch’s maximums averaged 8.2 degrees above what the regional stations imply. The day-to-day scatter in those weeks was more than twice the normal level, and in July 1918 it was more than three times as large.

A second, independent analysis points the same way. In April 2025 the Austrian meteorologist Reinhold Steinacker of the University of Vienna published his own study of the record in Oxford Open Climate Change, calling it “more than questionable” and putting the true value probably close to 125°F. He also estimated how likely the valley is to reach 130°F or more again in the decade from 2024 to 2033. Depending on the method, he got roughly 82 percent and 74 percent, but he considered a true 135 in the coming years unlikely.

The BAMS authors add one more simple argument. Death Valley has only recently reached 130°F, in 2020 and 2021, in a warmer modern climate, with modern instruments and careful siting. A week in 1913 that produced eight straight days of 127°F or more, in a cooler climate, from a shelter beside an irrigated field, is very hard to believe.

The record that came before

If the case against Greenland Ranch sounds familiar, it is because the world has been here before. For ninety years, 134°F was not even the world record. That title belonged to El Azizia, a town in Libya about 40 kilometers south-southwest of Tripoli, where a reading of 58°C, or 136.4°F, was recorded on September 13, 1922.

The WMO convened an international committee to examine it. The investigators identified five problems: a doubtful instrument, a new and inexperienced observer, a poor site, poor agreement with other stations in the region, and poor agreement with every later reading at El Azizia itself. The decisive evidence came from Libya’s own archives. A Libyan meteorologist, Khalid El Fadli, found the original logbook from 1922. It showed that a new observer had taken over on September 11, two days before the record, and that the minimum and maximum columns had been mixed up. The instrument was a Bellani-Six thermometer, a type that works much like the maximum-minimum thermometer in the old engraving above, with small sliding markers that record the extremes. The investigators concluded that the observer read the top of the marker instead of the bottom, an error worth about 7°C. His readings ran about 7°C above the four other stations in the area.

The investigation nearly stalled when Libya’s revolution broke out in 2011 and El Fadli dropped out of contact for a time; he resurfaced by email in August 2011. A reanalysis by the British climate scientist David Parker found the 1922 reading lay about six standard deviations from what the weather patterns of the day would allow. In January 2012 the committee voted unanimously to throw the El Azizia record out. “The eventual answer seemed so clear and obvious,” Tom Peterson, one of the scientists involved, said afterward. The world record passed, by default, to Greenland Ranch.

There is a nice irony in that. The same kind of forensic work that removed El Azizia, comparing one station to its neighbors, checking the observer and the instrument, rereading the original logs, is exactly what has now been applied to Death Valley. The WMO has also become much stricter about what it accepts. In 2019 it certified two of the hottest readings in modern times, 53.9°C at Mitribah, Kuwait, in July 2016 and 53.7°C at Turbat, Pakistan, in May 2017, only after the sensors were sent to an independent laboratory in Italy for calibration.

The modern 130s

Today the official Death Valley station sits near the park visitor center at Furnace Creek, about 1,000 feet north of the old Greenland Ranch site. It has been there since 1961, and it is now an automated sensor, installed in 2015, with the National Weather Service keeping watch.

On August 16, 2020, at 3:41 p.m., it recorded 130°F. If verified, the Park Service said, it would be the hottest temperature recorded on Earth in nearly a century. The new climate reference station at Stovepipe Wells, about 29 kilometers away and slightly higher, peaked at 124.7°F the same afternoon. On July 9, 2021, Furnace Creek hit 130 again. The park’s heat press kit gives the two readings as 129.9 and 130.1°F; the WMO lists both as 54.4°C.

Verifying them has taken years. In July 2023 the WMO said both the 2020 and 2021 sensors had to be dismantled and shipped for independent testing, and that one of the tests had been completed. At the same time, its rapporteur on weather and climate extremes, Randall Cerveny, made a point of saying the old record still stood. “Both the US National Extremes Committee and the WMO Archive … accept the 1913 temperature observation,” he said.

That is the sense in which the 2020 and 2021 readings matter to the 1913 argument. If 130°F is about as hot as Death Valley gets in a much warmer modern climate, measured carefully with tested equipment, then a 134 in 1913 needs an extraordinary explanation, and so far nobody has found one in the weather maps.

What happens to the record now

Changing a world weather record is not a matter of publishing a paper. The WMO keeps the global archive, but for a U.S. reading it leans on the National Climate Extremes Committee, the U.S. body that rules on national records. When the BAMS paper appeared online in the fall of 2025, Cerveny told the AMS news service that the WMO would defer to that committee, that the U.S. government shutdown then under way could slow things down, and that “the WMO will be in contact with the NCEC” given what he called “the purported new evidence.”

Reid, writing on his own weather blog in February 2026, described how the project came together: Spencer contacted him in November 2024 and soon asked him to join as a co-author. The paper ran in the January 2026 issue of the Bulletin. No ruling on the record had been announced in any of the sources reviewed for this story. Spencer has also suggested a practical step: a new climate reference station on the valley floor about seven miles south of Furnace Creek, since the existing one at Stovepipe Wells sits 278 feet higher than the official station, so that the next great heat wave is measured beyond argument.

Skeptics of the skeptics have a point too. A century-old reading should not be thrown out lightly, and Burt himself has always framed the question carefully. At the record’s centennial gathering in 2013, where people wore “134°” shirts, he talked about the danger of “propagating this myth of the 134 degrees,” but also said of old records generally, “They’re innocent until proven guilty.” The case against this one has been building for decades, with new evidence at every step. Whether it meets that standard is now up to a committee.

Old weather records are worth arguing over because they are almost the only window we have into the climate of the past. The cold summer of 1816, when a volcano brought snow to New England in June, is known from diaries, farm logs and the notes of careful observers who had no idea anyone would study them two centuries later. Most of those observers were honest and some of them were wrong, and sorting one from the other is the job.

The heat that needs no asterisk

Whatever happens to the 134, the valley’s current heat is not in doubt, and it is getting worse. July 2024 was the hottest month ever recorded in Death Valley, with an average temperature, day and night, of 108.5°F, beating the 108.1°F of July 2018. The average high that month was 121.9°F, nine days reached 125°F or more, and the peak was 129.2°F on July 7. “We just experienced the hottest month in history in the hottest place on Earth!” the park superintendent, Mike Reynolds, said.

The heat is dangerous in ways that the records do not show. On July 6, 2024, a group of six motorcyclists from Europe touring through the park became ill near Badwater Basin; one died of heat exposure and another was hospitalized. It was 128°F that afternoon, beating the day’s previous record of 127°F, set in 2007, and it was too hot for rescue helicopters, which generally cannot fly safely above about 115°F. That summer, which averaged 104.5°F, the park counted two heat-related deaths.

The park’s heat guidance is blunt. Water from a cold tap can come out above 105°F in summer. Above 120°F, routine outdoor work by park staff is banned, and emergency work is done in cycles of ten minutes on and fifty minutes off. The park’s advice to visitors fits on a sign: drink before you are thirsty. One ranger, Jennette Jurado, put the experience well: “Any wind feels like a hair dryer blowing on you.” Heat kills about one visitor every other year in the park, and August is often the park’s second-busiest month, as travelers come specifically to feel it.

A number on a T-shirt

In a way the 134 has become part of the landscape. It is on souvenirs and on the park’s own heat pages, and it is the reason many people make the drive in July. The park’s heat press kit still lists the old Greenland Ranch readings of 130, 131 and 134°F in July 1913 alongside the modern ones. That same year produced the valley’s coldest reading too, 15°F on January 8, which nobody has felt the need to challenge.

If the record falls, Death Valley will lose a line in the almanac but not its title. If the 2020 and 2021 readings are confirmed, they would be the hottest temperatures ever measured anywhere with modern, tested instruments, taken about 1,000 feet from where Denton kept his forms. The illusion would be gone, but the heat would be just as real.

What may change is the story we tell about it. For more than a century, the 134 has been offered as proof of what the desert can do. The newer story is about what people do with a number: a foreman alone in an oasis, sure that his thermometer was reading cold; a borax company that liked the sound of the hottest place on Earth; and scientists who, a hundred years later, went back to the paper and checked. The big thermometer outside the visitor center will keep posting the afternoon heat for visitors to photograph. These days, at least, someone is checking its arithmetic.

Further reading