The Hottest Temperature on Earth

by Brian Dunning
August 1, 2023

It's summertime in the age of climate change, and what does that mean? Heat waves, exceptional heat events, and records being broken. Today, one of those records in particular is going to receive the attention of our skeptical eye: it's the all-time highest temperature ever recorded on Earth. You've probably heard about it, Death Valley from the year 1913 — but have you also heard that it's almost certainly invalid?

I have a rather unusual disclaimer to give at the start of this episode, and that's that the information presented today may be already out of date. This may even happen by the time you hear it. We're discussing record high temperatures during a period in history when those records are being approached pretty regularly. Fortunately most of our discussion today is on older historical temps, a topic which is not impacted by any newer highest temperature we might record.

Today's topic is mainly an examination of what is — as of this writing — the official highest temperature recorded on Earth. The temperature used is the surface air temperature, which is taken over bare natural ground, 1.25-2 meters from the surface, and shielded from direct sunlight. At present, that official high is 134°F, taken at Greenland Ranch, Death Valley, California on July 10, 1913. (Only Fahrenheit was measured, and was recorded only in whole degrees.) It is almost certainly invalid, and is almost sure to be stricken from the record books, though it hasn't yet.

There actually is an official keeper of record temperatures, and that's the World Meteorological Organization, an agency of the United Nations. Their committees are composed of top meteorologists from nations all around the world. Run your eye down the list and it's experts from national meteorological services, top meteorological universities, and — one name sticks out a bit, Christopher Burt representing — who? Something called Weather Underground LLC. Amid nations and universities, some random company. Often portmanteaued into Wunderground, they sell weather reports over the Internet to outlets everywhere. Their name is taken from the far-left militant group that was founded at the same place as they were a quarter of a century earlier: the University of Michigan. Their data comes from the National Weather Service and also from a very unique source: over a quarter of a million personal weather stations that adhere to strict technical requirements. It's radical weather data, by the people, for the people. Wunderground's blog, called Category 6, is internally peer reviewed and truly is one of the world's best and most authoritative weather blogs. So when you see Wunderground rubbing shoulders with the world's top dogs at the WMO, it's very well earned.

This record high temperature makes it into the news every time Death Valley approaches it, and here's why that's especially important these days. The 134°F is brought up by climate change deniers, and flaunted as proof that we're not any hotter now than we were a century ago. This is not only factually wrong, it's invalid logic; random spikes in temperatures always happen and do not indicate overall trends, and we're unquestionably hotter than we were a century ago. But at least striking this record from the books will deprive the climate deniers of one more tool that can be compelling to people with only a layperson's understanding of the topic.

134°F became the WMO's official record in September 2012 when the previous record holder was deemed invalid by the WMO and was stricken. This was 58°C, also recorded only in whole numbers but equivalent to 136.4°F, in El Azizia in Libya in September 1922. On that 13 member committee that investigated the Libyan reading was Christopher Burt. Five basic reasons were given for striking the record:

1. Problematic instrumentation. The thermometer used was a replacement for the official instrument, and was of a household (non-scientific) type. It used alcohol along with mercury, and the alcohol could get between the mercury and the glass and give an invalid reading. It had a reset procedure that had to be done daily using a magnet.
   
   
2. New and inexperienced observer. The observer was a new guy who had started only two days before, and who was untrained on this type of thermometer and was likely unaware of its limitations. The committee's consensus was that an error of approximately 7°C was probable if viewed by an inexperienced observer.
   
   
3. Unrepresentative microclimate. The thermometer was located on a tarred concrete plaza, making the surface air temperature substantially higher than the surrounding desert. It was relocated to a compliant site in 1927.
   
   
4. Poor correspondence of the reading to other locations. Typically the station at El Azizia gives readings very close to others nearby, but for just the one month following the new observer's start, El Azizia recorded far higher temps.
   
   
5. Poor comparison to subsequent readings. The highs measured during that month did not correspond to the difference between highs and lows typically measured there, and were otherwise statistically out of whack.

So the El Azizia record didn't just have poor reliability, it had significant evidence that it was actually wrong. It was duly determined to be insufficiently reliable to retain the record, and it's now gone.

So now we turn to the next temperature standing, that famous 134°F (equivalent to 56.7°C) on July 10, 1913 at Death Valley. It still has the record, but it's increasingly noted as unreliable in news reports. A tweet by Scientific American in July 2023 warned that the coming weekend could crack 130°F, making it "the hottest temperature ever reliably measured on Earth". That set off a tweetstorm of criticism from climate deniers, accusing Scientific American of rewriting history to sweep the 1913 record under the rug, all to protect the global warming "narrative"; claiming that since the hottest temperature was way back in 1913 we're actually getting cooler, not warmer.

If we did decide to strike that current record, that would leave another Death Valley temperature as the new leader, which is 54.4°C/130.0°F, which has been recorded twice at Furnace Creek in August 2020 and again in July 2021. Here are the arguments for doing so.

It is Christopher Burt, once again, leading the movement to re-investigate the 1913 record and strike it from the books, and given his past work with the WMO, might well be on the committee to do so. In 2016 he wrote a lengthy article for Category 6 arguing his rationale, titled "An Investigation of Death Valley's 134°F World Temperature Record." From his summary:

This blog is courtesy of William T. Reid, a geographer and climatologist who has been studying the desert climate of California and, in particular, the Death Valley temperature record for some 30 years. Mr. Reid and I worked together to come to a commensurate conclusion regarding the validity of this significant planetary weather record: It is possible to demonstrate that a temperature of 134°F in Death Valley on July 10, 1913, was essentially not possible from a meteorological perspective, using an officially sanctioned USWB shelter and thermometer and following proper procedures observationally. Thus, the best explanation for the record high report(s) in July 1913 is observer error.

That observer was Oscar Denton, employed by the US Weather Bureau at Greenland Ranch, about 300 meters from the current station, and where at the time they grew alfalfa fields for the mules used to haul borax wagons. Denton's main job was as the ranch foreman working for the Pacific Coast Borax Company, a position he held for about eight years. 1913 was to be his first full summer with the side gig of weather observer. The fact that Denton was mainly a ranch guy and was not "scientifically inclined" — as Reid and Burt put it — may have played a role. Their lines of evidence that the 1913 record is invalid are three. Let's take them one at a time.

1. Greenland Ranch temperatures not consistent with meteorological conditions in July 1913

Extraordinarily high temps like 56.7°C/134°F require extraordinary heat events, namely, an exceptional heat wave. And the simple fact is that there was no heat wave, as demonstrated by normal temperatures having been recorded throughout the rest of the southwest. When this record reading was first reported, it was by a George Wilson writing in the Monthly Weather Review. He said, in part:

The condition was probably local as is often the case in mountainous regions, and the exceptionally high temperatures were confined to Death Valley.

Sounds reasonable — except Wilson was writing in 1913, the very dawn of the age of comprehensive meteorological data for the American west. Could Death Valley's extreme high have been isolated to just that tiny local area? Reid and Burt wrote:

We now know that this could not have been the case, since no heat event in Death Valley has ever, in over 100 years of observation, confirmed such a possibility.

For such a high temperature reading, the atmosphere above the region must be hot enough to support it, and the rest of the region shows that just wasn't the case.

2. Lack of correspondence with surrounding weather sites at time of July 1913 observations

This is a similar objection, but it's more specific. At the time, there were 10 other US Weather Bureau stations in the vicinity, ranging in altitude from Death Valley's below sea level to all the way up to just over 1800m. Temperatures are cooler the higher you go, and in that region, you could always plot all 11 of these stations on a graph with altitude on one axis and high temperature on the other, and they'd always form a nearly straight line. This has been consistently reliable throughout the history of these measurements.

Except for the one week in which Denton recorded the record temp, July 7-14, 1913, with highs ranging from 127°F to the record 134°F. It was a bit of a warm week; the other stations ranged from 7 to 11 °F higher than normal. Except for Denton's Death Valley station, which was 18°F hotter. On the graphs, his recorded numbers for the week were way off the line — something that's never happened before or since his time there.

Simply put, the divergence of the high temps recorded that week cannot be reconciled with the established meteorological history there.

3. Concerns with Denton's credibility and experience

Scientifically inclined or not, it should have been no great hardship for Denton to have read the thermometer and written down the reading. Except for one problem: there is evidence suggesting that Denton often did not do his job, but went back to the book later and just filled in any old number. Not just occasionally, but often.

For two weeks in 1914, Denton's book showed a high of either 109°F or 110°F, an odd string of consistency not seen in the records from nearby stations. Whenever it rained, he recorded the amount using one of only four numbers: 0.01", 0.1", 0.2", or 0.3". Some of the numbers he recorded indicated that he was resetting the maximum thermometer multiple times per day, in violation of procedure. During two whole winters, the minimum daily temperatures he recorded were much too warm, giving an implausibly narrow temperature range for two whole winters — indicating that he didn't know how to reset the minimum thermometer or wasn't bothering to. But these examples are the tip of the iceberg; Reid and Burt went on at length about various problems with the meteorological data Denton recorded — including that he apparently never once took a day off in all those years, like every observer was allowed to — not a single blank appears in all his records, further fueling the suspicion that he always filled in blanks with any random numbers that looked good to him.

Really, the best explanation for the 1913 reading of 56.7°C/134°F is observer error, as Reid and Burt put it, or as I would put it, "he took a week off and then filled in random numbers for the days he was gone."

Will the WMO re-investigate that record and strike it from the books? Well, we don't know their process; but if it does happen, we can be reasonably confident that it's for the best.

By Brian Dunning
Follow @BrianDunning