From Climate Etc.

By Mila Zinkova

Reassessing The Coldest March by Susan Solomon

On Thursday, 29 March 1912, Captain Scott (1914) made the final entry in his Journal. He wrote: 

Since the 21st we have had a continuous gale from W.S.W. and S.W. We had fuel to make two cups of tea apiece and bare food for two days on the 20th. Every day we have been ready to start for our depot 11 miles away, but outside the door of the tent it remains a scene of whirling drift. I do not think we can hope for any better things now. We shall stick it out to the end, but we are getting weaker, of course, and the end cannot be far.

It seems a pity, but I do not think I can write more.

R. SCOTT.

For God’s sake look after our people

A little over two months earlier (17 January 1912), Captain Scott and his companions arrived at the South Pole only to discover that the Norwegians were there first. The Norwegian party led by Roald Amundsen reached the South Pole on 14 December 1911. Extremely disappointed, Scott and his men began their return crossing on 19 January 1912.  It was a sad journey. They encountered unusually cold temperatures during March (Scott, 1914; Solomon, 2002). Two members of Scott’s polar party, Edgar Evans and Captain Lawrence Oates, perished before the party camped for the last time

Since 19 March 1912, Scott and two remaining members of the Polar Party (Doctor Edward Adrian Wilson and Henry Robertson Bowers) had been unable to resume their march due to a persisting blizzard. They were only 11 miles (less than 18 km) from the One Ton Depot that had a fresh supply of food and fuel

This Final Blizzard became a matter of controversy nearly a century later. Every so often, a scientific claim becomes so widely repeated that it hardens into “fact,” even when the evidence beneath it consists of… well… let’s just call it creative meteorology. Such is the case with the final blizzard in Captain Scott’s 1912 Antarctic tragedy. Many writers before and after Solomon published her book have claimed that Scott was dishonest or fabricated the weather, but none of them were prominent atmospheric scientists who actually worked in Antarctica. That is precisely why Solomon’s assertions carried unusual weight—and why their scientific flaws matter even more

In The Coldest March, Susan Solomon famously concludes that such a storm was “virtually impossible.” She writes:

Wilson and Bowers met their deaths with the injured Scott, but the scientific constraints of modern meteorology as shown here suggest that their deaths may have been a matter of choice rather than chance. Whether such a choice was made, and whether it reflected their own dedication or an order by a desperate Scott vainly attempting to save legacies rather than lives is a question not for science but for the human heart

This allegation has been widely echoed in popular media. Coverage in The New York Times (Chang, 2001), The Guardian (Glancey, 2001), and the Los Angeles Times (Hotz, 2001), all repeated the quote about “a desperate Scott vainly attempting to save legacies rather than lives”

But here’s the problem: the science doesn’t agree. Not modern satellite imagery, not historical data, not even the meteorological literature available before the book was published.

Zinkova (2025) has shown that the blizzard unfolded much as Scott described. Building on that foundation, this post turns to Solomon’s claims, focusing on the evidence that was overlooked or misinterpreted. 

Corner Camp: The New Oracle of Truth?

One of Solomon’s central arguments rests on the assumption that if Corner Camp didn’t have a blizzard, then Scott’s Last Camp couldn’t have had one either. To prove the claim, Solomon (2002 p. 318) presents Figure 68, captioned as follows:

Figure 68. Winds observed at the automated weather station near Last Camp and near Corner Camp (stations 4 and 3, respectively; see map 2) in March 1998. Note that the winds at the two locations closely track each other. The similarities shown have been observed in every strong March blizzard measured on the Barrier in the automated record (since 1984).

To evaluate this claim, I examined AWS data for March 1996—well within the 15-year period Solomon cites (1984–1998). As the graph below shows, the winds at the two sites do not track each other at all in this particular situation. In fact, the graph shows that on March 18, 19, 1996, during a strong blizzard at Scott’s Last Camp, Corner Camp remains nearly calm. And this isn’t an isolated anomaly: several similar cases appear in that same month alone.

In other words, the central premise of Solomon’s argument—“no storm at Corner Camp = Scott must have exaggerated”—collapses immediately when actual data are examined.

The Case of the Selectively Quoted Explorer

In the book, Solomon (2002, p. 311) cites a story from Cherry-Garrard’s The Worst Journey in the World:

The ferocious wind that spirited away the tent of the Cape Crozier party and threatened the survival of Wilson, Bowers, and Cherry-Garrard in July 1911 was a textbook example of a katabatic blizzard. It blew with hurricane force at more than seventy miles per hour, but the trio survived the storm in large part because it subsided in less than forty-eight hours, when the reservoir of cold air was emptied

However, in the same paragraph where Cherry-Garrard recounts his own blizzard experience, he also refers to the accounts of others, noting that “parties which had come to Cape Crozier in the spring had experienced blizzards which lasted eight or ten days”:

I knew that parties which had come to Cape Crozier in the spring had experienced blizzards which lasted eight or ten days. (Cherry-Garrard ,1922 p. 282)

The omission of the “eight or ten days” blizzards in Solomon (2002) is puzzling. Not at the same page but in the same book, Cherry-Garrard (1922 p. 446) describes yet another blizzard. It was an eight days long blizzard at Cape Evans:

“This blizzard lasted for eight days , up till then the longest blizzard we had experienced [at Cape Evans]: ” It died as it had lived , blowing hard to the last , averaging 68 miles an hour from the south , and then 56 miles an hour from the north finally back to the south , and so to calm.“

This record was not obscure. It was not lost. It was simply… not used. An eight-day blizzard at Cape Evans could easily have stretched to ten days or longer at Scott’s Last Camp. The reason is that barrier blizzards strike Cape Evans and the Last Camp simultaneously (Zinkova 2025). Yet when a barrier blizzard subsides, a katabatic blizzard often follows at the Last Camp (but not at Cape Evans), prolonging the storm’s impact (Zinkova, 2025)

A Decade Later

One might hope that, with time, the scientific errors in The Coldest March would fade away quietly. Instead, ten years after the book’s publication, Solomon (2012) repeated the same arguments in a public presentation—now with an additional meteorological flourish. While responding to questions (Solomon, 2012) declared:

The conditions on the barrier are pretty strongly katabatic. I mean, when you have this kind of blizzard, that’s normally what you’re getting, because you’re shielded so well from cyclones moving in from the, from the coast. So it’s a very different kind of meteorology. 

The conditions on the Ross Ice Shelf are not strongly katabatic at all, and in the region of Scott’s Last Camp, virtually every blizzard—katabatic, barrier, or cyclonic—is driven by synoptic-scale cyclones or mesoscale cyclonic activity.

It is unclear how this “shielded from cyclones” allegation came about because in her own book, Solomon (2002 p.350) cited the paper by Bromwich and Carrasco titled: Cyclonically forced barrier winds along the Transantarctic mountains near Ross Island.

To claim that the Ross Ice Shelf contradicts the very mechanisms that produce the winds Solomon attempted to analyze. In short, a decade later, the misunderstanding wasn’t corrected.
It was amplified.

The satellite perspective

If anyone still doubts how easily Solomon’s interpretation can be disproved, consider this: a single infrared satellite image is enough to show why her core assumption—that Corner Camp and Scott’s Last Camp always experience storms together—is scientifically untenable. Below is one example (April 15, 2018)

In this infrared frame, a katabatic blizzard is clearly pouring down from the plateau onto the Ross Ice Shelf, slamming directly into the region of Scott’s Last Camp (S). Meanwhile Corner Camp is calm and quiet. This kind of decoupling is routine, not exceptional.

And why does the katabatic flow show up in infrared? Because descending katabatic air warms adiabatically, producing a warmer signature on IR imagery compared to the frigid surface beneath it.

What else do we see in this very same image? A well-defined cyclone over the Ross Ice Shelf driving the synoptic setup. Without this cyclone katabatic winds would not have been able to reach the area of Scott’s Last Camp because they need pressure gradients to traverse the Ross Ice Shelf.  

I looked at hundreds of infrared satellite images to learn about every possible situation at the Ross Ice Shelf. On the other hand, Solomon (2002, 2012) relied on a single infrared satellite image—Figure 65 (Solomon, 2002, p.312)—as the basis for her broad meteorological conclusions and her remarkable speculation that Scott may have ordered Wilson and Bowers to die. Apparently drawing her conclusion from a single satellite image, Solomon (2002, p. 317) stated:

“The two locations [Scott’s Last Camp and Corner Cam] are inextricably linked by the basic physics of fluid flow: the southerly wind from the Barrier has no option but to continue on its course until it passes Corner Camp […] “

Yet even this lone satellite image reveals a cyclone, which Solomon seems not to have recognized at all.

Solomon’s misunderstanding of katabatic winds hadn’t shifted even a decade later. In 2012, she told her audience that the Ross Ice Shelf is essentially a flat barrier, and that the wind can only follow the flow because there’s nothing out there to break it up.

Solomon (2002, p. 310) quoted Captain Scott’s Message to the Public, where he rather plainly described “a continuous gale from W.S.W. and S.W.”? That wind direction points to a katabatic flow spilling from Byrd Glacier (Seefeldt et al., 2007).

And, more importantly, a wind coming from W.S.W. or S.W. at Scott’s Last Camp would not detour to Corner Camp, which sits almost directly south. In actuality, the direction and speed of katabatic winds are as variable as the weather systems that drive them, enabling them to traverse long distances across the Ross Ice Shelf. For example, the image below shows a situation in which neither of the two marked locations experienced a katabatic blizzard.

The following image indicates that Scott’s Last Camp was struck by a katabatic blizzard, while Corner Camp remained calm. Notice the cyclone

Solomon’s book did not reference several highly relevant papers on the Ross Ice Shelf airstream. For example, an article by Bromwich, Carrasco et al. (1993), in which the authors described a katabatic wind surge in July 1988 and specifically stated that every site recorded stronger-than-monthly-average winds… except Corner Camp, which apparently opted out of the excitement. Corner Camp (AWS 07 in their article) was the only place that “did not record stronger wind than its monthly average” (Bromwich, Carrasco et al.,1993).

In summary, this cascade of methodological failures in Solomon (2002) ultimately crystallized into a global media narrative that tarred the reputation of Captain Scott.

For those interested in what actually happened—and why the Final Blizzard likely transpired the way Captain Scott described it—my peer-reviewed analysis is available [here]. See also my Youtube [video.] Both explain in detail what the satellite data, the reanalysis, and the historical records actually show—namely, that the March 1912 storm was a straightforward product of synoptic-scale cyclonicic activity

References

Bromwich, D. H., Carrasco, J. F., Liu, Z., & Tzeng, R.-Y. (1993). Hemispheric atmospheric variations and oceanographic impacts associated with katabatic surges across the Ross Ice Shelf, Antarctica. Journal of Geophysical Research: Atmospheres, 98(D7), 13045–13062. https://doi.org/10.1029/93JD00879

Chang K. 2001. How Bad Luck Tipped the Scales to Disaster. New York Times. https://www.nytimes.com/2001/08/28/science/how-bad-luck-tipped-the-scales-to-disaster.html. (accessed on 20 August 2020).

Cherry-Garrard, A. (1922) The Worst Journey in the World, Antarctic, 1910-1913. Constable and Company Limited

Glancey, J. (2001). We could be heroes. The Guardian. https://www.theguardian.com/books/2001/nov/10/historybooks.highereducation1 (accessed 20 August 2020).

Hotz, R. L. (2001). Did Robert Falcon Scott tell the truth? Los Angeles Times. https://www.latimes.com/archives/la-xpm-2001-dec-23-bk-hotz23-story.html (accessed 20 August 2020).

Scott R. F. (1914).Scott’s Last Expedition. Smith, Elder and Company, London.

Seefeldt M.W., Cassano J.J., and Parish T. R. (2007). Dominant Regimes of the Ross Ice Shelf Surface Wind Field during Austral Autumn 2005. J. Appl. Meteor. Climatol., 46: 1933–1955, https://doi.org/10.1175/2007JAMC1442.1.

Solomon, S. (2002). The Coldest March: Scott’s Fatal Antarctic Expedition. Yale University Press.

Solomon, S. (2012). The Coldest March: Scott’s Fatal Antarctic Expedition. https://youtu.be/5LoWsLqcizA (accessed 10 January 2022).

Zinkova, M. (2025). An exceptional March 1912 blizzard (the Final Blizzard) that sealed the fate of Captain Scott and his party. Arctic, Antarctic, and Alpine Research, 57(1), 2522490. https://doi.org/10.1080/15230430.2025.2522490