World Heat Record Overturned--A Personal Account

World Heat Record Overturned--A Personal Account

Almost two years after my October 8, 2010 wunderground blog post, Questions Concerning the World Record Temperature of 58°C (136.4°F) at Al Azizia, Libya on September 13, 1922, the World Meteorological Organization (WMO) has come to the conclusion that the record, in fact, was invalid. Here is a personal account of how this decision came to be. The narrative below is adapted from a blog post I have written for the Bulletin of the American Meteorological Society (BAMS).


A photograph of the trading post at Al Azizia, Libya taken in 1923. The photo was taken from the Italian military fort located on a small hill just south of the trading post. It was at this fort that the temperature of 58°C (136.4°F) was observed on Sept. 13, 1922 (Used with permission from the family of Gen. Enrico Pezzi).

As any weather aficionado can avow, Earth's most iconic weather record has long been the legendary all-time hottest temperature of 58°C (136.4°F) measured on September 13, 1922 at Al Azizia, Libya (also called El Azizia, there are many variant spellings). It is a figure that has been for meteorologists as Mt. Everest is for geographers. For the past 90 years, no place on Earth has come close to beating this reading from Al Azizia, and for good reason--the record is simply not believable.

In early March 2010, I was included in an email loop concerning questions about this record. The email discussion participants at that time included Maximiliano Herrera, an Italian temperature researcher and climatologist based in Bangkok, Piotr Djakow, a Polish weather researcher, and Khalid Ibrahim El Fadli, director of the climate department at the Libyan National Meteorological Center (LNMC) in Tripoli.

Previous to this discussion, I had generally accepted the Libyan world record as acceptable, although suspicious. The figure had been around for 90 years, and two previous studies by Amicare Fantoli (who was the man responsible for verifying the record in 1922) had more or less substantiated the extreme 58°C figure.

However, Piotr produced a chart of the monthly temperature amplitudes at Azizia for each September from 1921 - 1940, and this chart raised an alarm so far as the validity of the Aziza record was concerned. This was the first time that I began to really think something was not right about the record.


A chart showing the average monthly temperature amplitudes (difference between daily minimum and daily maximum temperatures) for Al Azizia during the month of September from 1919 - 1940. In 1927, the station was moved from the military fort on the hill to the town below and placed in civilian hands. (Chart produced by Piotr Djakow).

In September 2010, Weather Underground hired me as their Weather Historian, proposing that I write a weekly blog on extreme weather events and records. I decided that one of my first blogs should on the Al Azizia record (in fact it was the third blog I wrote for WU).

I was intrigued that El Fadli was skeptical of the Al Azizia 58°C figure, and requested more data. El Fadli’s enthusiastic and gracious response (to provide all and any weather data I might be interested in) was beyond my expectations. Past experience had shown me that many national weather bureaus consider their data proprietary and/or subject to excessive fees for access.

With El Fadli’s data on hand and after researching all (to me at the time) known other references concerning the Al Azizia event, I posted a blog on wunderground.com reflecting my findings on October 8, 2010. I forwarded a copy of this to Dr. Randy Cerveny, a professor at Arizona State University (ASU) and co-Rapporteur of climate and weather extremes for the World Meteorological Organization (WMO).

It should be noted that without the credibility of Weather Underground as my sponsor, Dr. Cerveny probably would not have taken the blog seriously. After all, many people in the past have questioned the validity of this record both in published works and on the Internet.

In any case, Randy picked up the ball and created an ad-hoc evaluation committee for the World Meteorological Organization to evaluate the record for the WMO Archive of Weather and Climate Extremes (http://wmo.asu.edu/). After this positive response from Randy, I asked El Fadli if Libya officially accepted the Azizia figure. He responded that they did not. Since records like this are, to a degree, the provenance of national interest and El Fadli responded that Libya did not officially accept the colonial-era data from Azizia (measured by Italian authorities at that time in Tripolitania), this became the catalyst to launch an official WMO investigation.

This would be an unprecedented investigation for this WMO extreme records evaluation committee. Rehashing old records is not the WMO Archive’s primary objective, which is to verify new potential records. As Dr. Tom Peterson of the US National Climate Data Center and President of the WMO’s Commission on Climatology (of which the Archive is a part), put it:

“To be honest, I was reluctant to reopen this question because other people had looked at the record in the past and it had been so widely accepted. I was particularly afraid that it would be an uncertain subjective opinion as to whether it was a bit off or not.”

Nevertheless, the investigation was approved and on February 8, 2011 an international team of climate experts was assembled (eventually 13 atmospheric scientists in all) by Randy. The official investigation began.

Amazingly, El Fadli had just uncovered a key document: the actual log sheet of the observations made at Azizia in September 1922 (see illustration further below). The log sheet clearly illustrated that a change of observers had occurred (as was evidenced by the hand written script) on September 11, 1922, just two days prior to the ostensible record temperature of 58° on September 13th. Furthermore, the new observer had interchanged the Tmin columns with the Tmax columns.


A copy of the log sheet from El Azizia for September 1922 found by El Fadli in January 2011. Image courtesy of the Libyan National Meteorological Center (LNMC).

Also, beginning on September 11th the Azizia maximum daily temperature records began to exceed by 7°C, on average, the other four stations reporting from northwest Libya (Tripolitania) at that time. That trend continued for the rest of the month (with a couple of days of missing data), and into October 1922.


A graph comparing the daily maximum temperatures observed in September 1922 at the five meteorological stations that existed in Tripolitania (northwestern Libya) at that time. There was an observer change at El Azizia beginning September 11th. (Graphic produced by Jim Petit).

Just as this key discovery (the finding of the original log sheet) was made, the Libyan revolution broke out. On February 15, 2011, we received the last message from El Fadli prior to the revolution. Col. Gaddafi, the leader of Libya, had shut down Libyan international communications.

Of course, without El Fadli’s critical input we could move no further with the investigation, and Randy called for a hiatus to further deliberations.

In early March, Gaddafi began airing long nightly rambling tirades on his government TV network. During one of these, he made an ominous reference to how NATO forces were using Libyan climate data to plan their assault on the country. My heart sank when I heard this. I immediately thought that our colleague, El Fadli--as director of the LNMC--must have been implicated by Gaddafi as providing weather information to the "enemy".

I must say, at that point, I--and the rest of the committee--thought El Fadli was a dead man.

We didn’t hear again from El Fadli until August 2011 when the revolutionary forces closed in on Tripoli. One of our committee members, Dr. Manola Brunet (WMO chair of the Open Programme Area Group on Monitoring and Analysis of Climate Variability and Change), who knew El Fadli personally, had up until then been unable to contact him by phone or email. Then on August 13, 2011, we received our first email from El Fadli.

El Fadli here relates the situation he faced during those long months when we lost communication with him:

“During that critical time all communication systems in Libya were shut down by the regime so it was impossible to communicate with anyone, even inside the country. Mobile telephone communications were restricted and even local calls were controlled and monitored. What was amazing however, believe or not, was that my office satellite Internet connection was still up and running. But using such posed serious dangers, if anyone discovered me I would probably lose my life. Hence, I never used that connection. The first 3 months (February-May) I was able to reach my office (my home being about 5 km east of El Azizia and 40 km to my office in Tripoli) but then in May we suffered from short fuel supplies, electricity, and even cooking gas. You can imagine how difficult our lives became! The other serious story involved the security situation. When I borrowed a car belonging to the local United Nations office (since I had no fuel for my own car) I was driving to morning prayers (04:00 am) with my sons and suddenly we came under gunfire from the back and rear of the vehicle. The vehicle was struck as I drove at a crazy speed with our heads ducked low. Our life was spared by the grace of God. This happened in late July.”

Then, as we all watched through the technology of television and Internet, by September 2011, the dictator Gaddafi was gone … and El Fadli was back!

With the investigation back on track, committee members made further progress in October and November. Dr. David Parker of the U.K. Met Office did a reanalysis of surface conditions across the Libyan region for September 1922. The results displayed a significant departure (up to 6 sigmas) from what the temperature observed at Azizia was to what the reanalysis plotted for the area. This was a key discovery, using technology that had never been available in past investigations of the Libyan record.



A chart produced by committee member David Parker following his reanalysis of the Azizia record. One could clearly see how the maximum temperature of 58°C in September 1922 was well beyond the top percentile that might have been expected.

Also, Philip Eden of the Royal Meteorological Society and others uncovered information concerning the unreliability of the Bellani-Six type of thermometer that had apparently been used at Azizia in September 1922. Of particular interest was how the slide within the thermometer casing was of a length equivalent to 7°C.  It would be easy for an inexperienced observer to mistakenly read the top of the slide for the daily maximum temperature rather than correctly reading the bottom of such slide, a point that El Fadli made in a message to me early on in the investigation.


A 1933 instrument catalog image of the Bellani-Six style thermometer. Image supplied by Paolo Brenni, President of the Scientific Instrument Commission, and courtesy of Library of the Observatorio Astronomico Di Palermo, Gisuseppe S. Vaiana.

With all the pieces of the puzzle now falling into place, a vote was taken in January 2012 resulting in a unanimous decision by the WMO committee members to disallow the Azizia record.

As Tom Peterson put it, “The eventual answer seemed so clear and obvious that we evidently must have done a far more in depth investigation than any earlier one.”

Conclusions
In the end, and based on the unanimous decision by the 13 committee members, Randy Cerveny and Jose Luis Stella of Argentina, (the WMO’s co-Rapporteurs of climate and weather extremes), rejected the 58ºC temperature extreme measured at El Azizia in 1922.

The WMO committee added the following comment to my blog as it appeared on the BAMS web site front page: ”An important aspect of this long investigation was that it just isn’t climatologists and meteorologists changing their minds. It goes beyond that. This investigation demonstrates that, because of continued improvements in meteorology and climatology, researchers can now reanalyze past weather records in much more detail and with greater precision than ever before. The end result is an even better set of data for analysis of important global and regional questions involving climate change. Additionally, it shows the effectiveness of truly global cooperation and analysis. Consequently, the WMO assessment is that the official highest recorded surface temperature of 56.7°C (134°F) was measured on 10 July 1913 at Greenland Ranch (Death Valley) CA USA.

I agree with this comment as well, although I still have my suspicions about the Greenland Ranch figure of 1913. Time for another investigation!

REFERENCE NOTE:For the official WMO assessment report and all the references associated therewith please see the September 2012 issue of BAMS magazine (Bulletin of the American Meteorological Society).

KUDOS: Maximiliano Herrera for initially bringing the Azizia issue to my attention, Piotr Djakow for producing the first graphic that made the Azizia error so clear, Jim Pettit for helping me along the way with graphic analysis and encouraging advise, and Randy Cerveny for picking the ball up and championing the investigation. Also, of course, a big thanks to El Fadli without whom this investigation would never have happened, all the other WMO Committee members for their diligent work. In addition I want to thank Weather Underground, especially Jeff Masters and Shaun Tanner, for providing me the platform from which I can blow my horn!

Links:
Don't miss the 25-minute wunderground video,Dead Heat, a detective story on how the Al Azizia record was overturned.

Christopher C. Burt
Weather Historian