The accounts of the eyewitnesses

The blizzard that stopped Scott's party at the RIS started on 5 December 1911, when Scott's party was ~12 miles from Beardmore Glacier (83°20′ S; Fig. 1) on their way south. However, the bad weather began earlier, on 2 December 1911. At that time, the challenging conditions were variable, and the party was still able to march south. Below are the descriptions of the weather conditions that the explorers from different parties encountered from 2 to 5 December 1911.

At the same time the Polar Party was experiencing the blizzard, another part of the British expedition (the Motor Party) was located at 79.31 S, 169.22 E (Fig. 1). Their weather observations are provided in Simpson (Reference Simpson1923: 661):

On 5 December 1911, the Norwegian party was at the Polar Plateau at ~88° S (and later 88°9′ S, 170° W). Amundsen (Reference Amundsen1913: 107) wrote in his journal:

On 6 December 1911, Amundsen recorded the temperature to be > -16°C (~3°F), ‘which represents an anomaly relative to our estimate from ERA-Interim climatology of more than 10° C’ (Fogt et al. Reference Fogt, Jones, Solomon, Jones and Goergens2017).

By 9 December 1911, 18–24 inches (457–610 mm) of snow was recorded at Cape Evans (Fig. 1; Simpson Reference Simpson1923: 306). Was this blizzard an EPE? Turner et al. (Reference Turner, Phillips, Thamban, Rahaman, Marshall and Wille2019) define an EPE as an event in which daily precipitation exceeds a percentile value of the 1979–2016 distribution (the 90th percentile in their study) for 1 or more days. The 90th percentile (mm) for the RIS is 1.16 mm (Turner et al. Reference Turner, Phillips, Thamban, Rahaman, Marshall and Wille2019). Thus, the snowfall the Polar Party endured on the RIS from 2 to 8 December 1911 was definitely an EPE, as evident from the weather conditions described by the eyewitnesses.

Did some parts of the RIS and the Transantarctic Mountains melt and refreeze in December 1911?

Several accounts by Scott (Reference Scott1913) appear to suggest that melting of some parts of the RIS and the Transantarctic Mountains did occur. After the storm subsided, the Polar Party entered the glacier's valley in the Transantarctic Mountains. Day after day, they continued to encounter very soft snow, in which ‘the sledges themselves sank to the crossbars in soft spots’ (Scott Reference Scott1913: 341). On 16 December 1911, Scott (Reference Scott1913: 348) wrote that ‘the snow had become wet and sticky'. The air temperature on that day was +15°F. On 19 December 1911, at a height of ~5800 feet (1768 m), Scott (Reference Scott1913: 351) fell in one of the ‘very annoying criss-cross cracks'. A few times Scott used the word ‘neve’ to describe the surface:

• • 15 December 1911: ‘patches of ice and hard neve are showing through in places'.

• • 19 December 1911: ‘The last mile, neve predominating'.

• • 23 December 1911: ‘we got on the most extraordinary surface - narrow crevasses ran in all directions. They were quite invisible, being covered with a thin crust of hardened neve without a sign of a crack in it.’ According to dead reckoning, the party was located at 85°22′1″ S., 159°31′ E. Their elevation was 7750 feet (2362 m; Scott Reference Scott1913: 355).

• • 25 December 1911: ‘to our annoyance found ourselves amongst crevasses once more - very hard, smooth neve between high ridges'. Dead reckoning was 85°50′ S, 159°8′2″ E (Scott Reference Scott1913: 357).

Miller (Reference Miller1952: 150) talked to several French and Swiss glaciologists (because névé is a French word) and they ‘consider that névé is a more or less dense and settled, although permeable, aggregate of medium to large individual grains formed and welded together by frequent alternations of melting and freezing on original snow crystals'.

Some melting and refreezing affected the RIS as well. On 27 February 1912, while going north over the RIS, Scott (Reference Scott1913: 401) encountered ‘[v]ery curious surface - soft recent sastrugi which sink underfoot, and between, a sort of flaky crust with large crystals beneath'. The author of this article asked a retired NASA scientist, Dr James L. Foster (who has published multiple papers on snowmelt), what he makes of Scott's description. Dr Foster responded: ‘Yes, to me, the mention of flaky crust seems to indicate some sort of melt and refreeze.'

The Twentieth Century Reanalysis

The Twentieth Century Reanalysis cannot be used to analyse the climatology of the December 1911 EPE. The key synoptic parameters (the anomaly of columnar precipitable water and 2 m mean air temperature) plotted with the Twentieth Century Reanalysis for the key days in December 1911 (Fig. 2a,b) do not agree with the descriptions provided by the members of the Terra Nova expeditions. The anomaly of columnar precipitable water (Fig. 2a) is negative, and the 2 m mean air temperature of 260 K (8°F) is much lower than that reported by Scott's party. This confirms Bromwich et al.'s (Reference Bromwich, Fogt, Hodges and Walsh2007) findings that early Antarctic reanalysis is unreliable and cannot be used to determine the climatology of the December 1911 EPE.

Figure 2. The plots created with the Twentieth Century Reanalysis. a. Anomaly of columnar precipitable water from 5 to 8 December 1911. b. The 2 m mean air temperature from 5 to 8 December 1911. c. Anomaly of columnar precipitable water from 6 to 8 January 2005. d. The 2 m mean air temperature from 6 to 8 January 2005. The position of Scott's Party is marked with the black ‘X'.

On the other hand, the synoptic parameters of a similar event that occurred in January 2005 (Fig. 2c,d) mostly agree with the conditions of the December 1911 EPE. Is it possible to use reanalysis plots and satellite images of th|e January 2005 event to better understand the climatology of the December 1911 event?

January 2005 event

Fogt (Reference Fogt2005) describes ‘a persistent low pressure and snow’ from 4 to 11 January 2005. The storm ‘disrupted all of the north-south flights during its stay, and many of the intercontinental operations'. Some of the roads at Scott's Base (New Zealand research station at Ross Island) were at condition 1, the worst weather condition possible, whereas McMurdo Station (the US research station at Ross Island; Fig. 1) was at condition 2 and sometimes, during a ‘several windows of good weather', at condition 3. The storm was a synoptic-scale cyclone that ‘remained in this location for several days due to ridging [a blocking high] over Marie Byrd Land'. Figure 3a (500 mb geopotential height anomaly) shows a blocking high that kept a synoptic-scale cyclone (Fig. 3b) in place in January 2005.

Figure 3. a. 500 mb geopotential height anomaly shows a blocking high over Marie Byrd Land (Fig. 1). b. Pressure at mean sea level shows a synoptic-scale cyclone affecting the Ross Ice Shelf.

Zou et al. (Reference Zou, Bromwich, Montenegro, Wang and Bai2021) identified January 2005 as a year with a ‘powerful moisture transport (~100 kg⋅m^-1⋅s^-1) from the Amundsen/Ross Sea region’ to the RIS, which expanded towards the Transantarctic Mountains. In 2005, moisture flux occurred from 3 to 7 January (Zou et al. Reference Zou, Bromwich, Montenegro, Wang and Bai2021). Two atmospheric rivers (ARs) made landfall at the RIS on 4 January (Fig. 4) and 6 January (Wille et al. Reference Wille, Favier, Dufour, Gorodetskaya, Turner, Agosta and Codron2019). The Glossary of Meteorology defines an AR as a ‘long, narrow, and transient corridor of strong horizontal water vapor transport that is typically associated with a low-level jet stream ahead of the cold front of an extratropical cyclone'. A satellite image of the AR that made landfall on the RIS on 3–4 January 2005 is presented in Fig. 4. The AR originated over mid-latitudes and transported moisture to Antarctica, embedded in the warm conveyor belt of an extratropical cyclone over the Ross Sea.

Figure 4. A Moderate Resolution Imaging Spectroradiometer (MODIS) satellite image of the Antarctic and sub-Antarctic Oceans for 3–4 January 2005. The MODIS Cloud Water Path indicates the amount of water in the atmosphere above a unit surface area of the Earth. The image reveals an atmospheric river, which transported moisture from the mid- to high latitudes, where it embedded itself into the extratropical cyclone (marked with the red ‘L') in the Ross Sea. Notice that the moisture advanced into the Ross Ice Shelf alongside the Transantarctic Mountains, to the same places, in which the members of the Terra Nova expedition endured the December 1911 4 day blizzard. Credit: NASA.

The January 2005 event resulted in above-normal air temperatures (Fig. 2d) and above-normal precipitation (Fig. 5a). According to the ERA5 plot (Fig. 5b), the winds along the Transantarctic Mountains were not affected, which probably means that the event of December 1911 was stronger.

Figure 5. The reanalysis plots created with ERA5 reveal the anomalies in a. precipitation and b. wind speed from 6 to 8 January 2005. The position of Scott's party is marked with the black ‘X'.

Not only can ARs cause EPEs, they also can cause severe melting over the RIS (Zou et al. Reference Zou, Bromwich, Montenegro, Wang and Bai2021) that can last for weeks (Scott et al. Reference Scott, Nicolas, Bromwich, Norris and Lubin2019). The meltwater can penetrate beneath the surface, and then, when the air temperature drops, this meltwater turns into ice. In January 2005, the surface melting on the RIS was captured by NASA satellites (Fig. 6), and the mechanisms of the melting were described by Wille et al. (Reference Wille, Favier, Dufour, Gorodetskaya, Turner, Agosta and Codron2019) and Zou et al. (Reference Zou, Bromwich, Montenegro, Wang and Bai2021). The 2005 melting event was rather unusual because it affected high elevations in the Transantarctic Mountains. Tedesco et al. (Reference Tedesco, Abdalati and Zwally2007) write: ‘Also, first-time [for the study years 1987–2006] extensive melting occurred over the Transantarctic Mountains and on the edge of the Ross Ice Shelf for the week between January 6–12, 2005, with melting reaching areas located 875 km inland and 2000 m above sea level.’ Wille et al. (Reference Wille, Favier, Dufour, Gorodetskaya, Turner, Agosta and Codron2019) attribute ‘nearly 100%’ of the melting events at the higher elevations in the Marie Byrd Land (Fig. 1) to the ARs. Although the authors do not mention the higher elevations in the Transantarctic Mountains, it is reasonable to assume that the tendency at the higher elevations in Marie Byrd Land and at the higher elevations in the Transantarctic Mountains would be approximately the same.

Figure 6. Satellite imagery shows the number of Antarctic melting days for the 2004–2005 season. Areas where melting occurred for a greater number of days are indicated in increasingly dark purple shading. Note the melting patches at > 85°S. An approximate location of Scott's party is marked with the black ‘S'. Credit: NASA/Rob Simmon.

Are there some similarities between the December 1911 event and the January 2005 event?

We have already proven that the December 1911 event was an EPE. Adusumilli et al. (Reference Adusumilli, Fish, Fricker and Medley2021) established that 63% of the EPEs over the West Antarctic Ice Sheet were associated with ARs. Wille et al. (Reference Wille, Favier, Gorodetskaya, Agosta, Kittel and Beeman2021) concluded that ‘70% of the highest 1% of precipitation events can be attributed to ARs across East Antarctica'. As we discussed earlier, the Norwegian party experienced an EPE on the Polar Plateau (East Antarctica) for > 2 days and a 10°C positive air temperature anomaly for 4 days (Fogt et al. Reference Fogt, Jones, Solomon, Jones and Goergens2017), which both suggest the intrusion of warm marine air.

Was the December 1911 EPE caused by the ARs? A few factors suggest that it was:

1. 1) The apparent melting over some parts of the western RIS and most importantly in the higher elevations in the Transantarctic Mountains as described by Scott (Reference Scott1913).

2. 2) The Southern Annular Mode (SAM) index for December 1911 was strongly negative (-3.867; Fogt et al. Reference Fogt, Jones, Solomon, Jones and Goergens2017). Negative SAM favours ARs affecting Marie Byrd Land and the RIS (Fig. 1; Wille et al. Reference Wille, Favier, Dufour, Gorodetskaya, Turner, Agosta and Codron2019).

3. 3) The duration of the event from 2 to 8 December 1911 strongly indicates the presence of a blocking high, and as Wille et al. (Reference Wille, Favier, Dufour, Gorodetskaya, Turner, Agosta and Codron2019) write: ‘AR landfalls are most likely when the circumpolar jet is highly amplified during blocking conditions in the Southern Ocean.'

4. 4) The significant increase in the air temperatures, not only on the RIS but also on the high Plateau, is consistent with intrusion of very warm marine air.

5. 5) Figure 3 reveals that the locations where the observations were made in December 1911, which are located in a very close proximity to the path that the ARs took in January 2005. In addition, the apparent refreezing that Scott (Reference Scott1913) observed in the Transantarctic Mountains at 85°22′1″ S., 159°31′ E is located alongside the melting-refreezing site of January 2005 (Fig. 6).