A Tale of Two Climbers: Hypothermia,Death,and Survival on Mount Everest

Abstract

Moore, G.W.K., J.L. Semple. A tale of two climbers: hypothermia, death, and survival on Mount Everest. High Alt. Med. Biol. 13:51–56.—Hypothermia is an acknowledged risk for those who venture into high altitude regions. There is however little quantitative information on this risk that can be used to implement mitigation strategies. Here we provide an analysis of the meteorological and hypothermic risk parameters, wind chill temperature, and facial frostbite time, during the spring 2006 Mount Everest climbing season. This season was marked by two high profile events where a solo climber was forced to spend the night in highly exposed conditions near the summit. One climber survived, while the other did not. Although this retrospective examination of two individual cases has admittedly a small sample size, and there are other factors that undoubtedly contributed to the difference in outcomes, we show that wind chill temperature and facial frostbite time experienced by the two climbers were dramatically different. In particular, the climber who did not survive experienced conditions that were approximately one standard deviation more severe that usual for that time of the year; while the climber who survived experienced conditions that were approximately one standard deviation less severe then usual. This suggests that the environmental conditions associated with hypothermia played an important role in the outcomes. This report confirms the importance of providing quantitative guidance to climbers as the risk of hypothermia on high mountains.

Introduction

The combination of low temperatures and high winds that are common in high altitude environments can compound hypoxic physiological stresses, resulting in an increased risk of hypothermia and frostbite (Huey and Eguskitza, 2001; Pandolf and Burr, 2002; Ward, 1974). A comprehensive analysis of mortality on Mount Everest indicated that hypothermia, along with high altitude illness, was the leading cause of nontraumatic fatalities amongst climbers (Firth et al., 2008). Towards this end, adverse weather conditions were a contributing factor in approximately 25% of the fatalities above 7000 m (Firth et al., 2008). The analysis of two high profile storms on Mount Everest that resulted in fatalities, the 1924 Mallory and Irvine summit attempt and the 1996 ‘Into Thin Air’, reported that physiologically significant drops in barometric pressure occurred during these events that were associated with high winds and cold temperatures (Moore and Semple, 2006; Moore et al., 2010). Anecdotal evidence also suggests that frostbite is also common among Sherpas who work in the Himalaya and for whom such cold injuries are an occupational health issue related to inadequate information and clothing (Subedi et al., 2010).

The risk of hypothermia is also significant in other regions. On Denali, the highest mountain in North America, hypothermia is the second leading cause of death, with 16% of all fatalities being attributed to it (McIntosh et al., 2008). On Mount Logan, North America's second highest mountain, a severe storm in May 2005, characterized by winds in excess of 30 m/s and temperatures of −15°C, trapped three members of an elite mountain rescue team at 5400 m for over 3 days (Bjarnason and Jardine, 2005; D'Oro, 2005; Moore and Holdsworth, 2007). All three members suffered frostbite injuries that in the case of one climber resulted in amputation of all fingers (Bjarnason and Jardine, 2005). In Iran, the incidence rate for frostbite among mountaineers is approximately 36%, with a third of these injuries occurring at elevations above 4400 m (Harirchi et al., 2005). It has also been reported that strong winds and cold temperatures are often associated with deaths in the Pyrenees (Pascual and Callado, 2010).

The first quantitative estimate of the risk of cold injury at high altitude has only recently been provided (Moore and Semple, 2011). This study focused on two parameters, the wind chill temperature (WCT), defined as the temperature in still air that would result in the same steady state facial heat loss as occurs at a given temperature and wind speed, and the facial frostbite time (FFT), defined as the time it takes facial flesh to freeze. Throughout the year near the summit of Mount Everest, it was shown that the typical WCTs and FFTs are always less −30°C and 20 min, respectively (Moore and Semple, 2011). During the spring climbing season, WCTs of −50°C and FFTs of 5 min are typical; while during severe storms they can approach −60°C and one minute, respectively. It was furthermore argued that the barometric pressure is an excellent predictor of both WCT and FFT, allowing for the possibility of using this relatively simple, robust, and easy to observe meteorological parameter as a predictor of the risk of hypothermia (Moore and Semple, 2011).

The 2006 spring climbing season on Mount Everest was one of the most deadly and controversial in history (Hall, 2008; Heil, 2008; Kodas, 2008). Two events in particular galvanized both the climbing community and public opinion as to the challenges, risks, and changing ethos associated with attempting to summit the mountain. On May 14, a solo climber, David Sharp from the United Kingdom, ran into distress as he descended late in the day from the summit region and was forced to bivouac at 8500 m (Heil, 2008). During the early hours on May 15, over 40 climbers passed his comatose body, some observing that he was still alive. However, no one stopped to provide assistance (Heil, 2008). By the time a rescue attempt was organized, Sharp was so hypothermic that he was unable to respond to supplemental oxygen and died sometime late on May 15.

Ten days later on May 25, an Australian climber, Lincoln Hall together with his Sherpas, summited Mount Everest (Hall, 2008; Heil, 2008). During the descent, Hall became disoriented and ultimately collapsed just below the Third Step at 8670 m. Sherpas stayed with him for as long as possible, before placing him in a relatively sheltered location and descending to safety. When they reached the North Col, it was reported that Hall had died (Hall, 2008; Heil, 2008). Early the next morning, Hall was found alive and coherent by another climbing party; who abandoned their summit attempt and were able to get him into warm clothing and provided oxygen. Through the efforts of a number of climbers and Sherpas, Hall was assisted back to the North Col where he received treatment for his severe frostbite and ultimately made a full recovery (Hall, 2008; Heil, 2008).

The very different outcomes from these two seemingly similar events, a solo bivouac above 8500 m by a disoriented hypoxic climber, raises the question as to why? In this report, we attempt to provide a quantitative answer through the analysis of the meteorological and hypothermic risk parameters associated with these two events. As we shall show, Sharp was exposed to WCT and FFT that were approximately one standard below the mean; while in Hall's case they were approximately one standard deviation above the mean. This marked difference most likely contributed to the dramatically different outcomes.

Methods

Research into the quantifying of the combined impact of wind and low temperatures have on the cooling of exposed skin has been ongoing for over 60 years (Osczevski and Bluestein, 2005; Siple and Passel, 1945). Joint work by the American and Canadian Weather Services has recently resulted in the development of new expressions for WCT and FFT that combine results from wind tunnel studies of heat loss from human subjects with model studies of the heat transfer from the human face (Office of the Federal Coordinator for Meteorological Research, 2003). These new expressions are currently being used for forecasting purposes in the two countries (Osczevski and Bluestein, 2005). It should be noted that both parameters are highly nonlinear functions of temperature and wind speed (Moore and Semple, 2011; Office of the Federal Coordinator for Meteorological Research, 2003). We also introduce a new parameter, the wind chill temperature deficit (WCTD), which is defined as the difference between the observed temperature and the wind chill temperature. As such, it provides an estimate of the additional cooling that is occurring as a result the combined effects of low temperatures and high wind speeds.

It should be noted that these expressions were derived at sea level and there are meteorological and physiological issues with applying them at altitude (Huey and Eguskitza, 2001; Moore and Semple, 2011). In particular, the convective heat loss is proportional to the square root of the Reynolds Number, which in turn depends on the product of the air density and wind speed (Tikuisis and Osczevski, 2002). Although there is a decrease in air density with altitude (Steadman, 1979), there is also an increase in wind speed that results from turbulent boundary flow that exists in regions of complex topography such as mountainous regions that is not captured in global meteorological datasets (Ahmad and Boraas, 2010; Moore and Semple, 2011). To the lowest order, these effects cancel, suggesting that these expressions represent valid estimates as to the risk of cold injury at high altitude (Moore and Semple, 2011). In addition, the physiological effects of high altitude including hypoxia and dehydration can increase the risk of cold injury and are not incorporated into the expressions (Ainslie and Reilly, 2003; Ward, 1974).

We use the Interim Reanalysis from the European Center for Medium Range Weather Forecasting-ERAI (Dee et al., 2011) to derive time series of wind speed and temperature at the mountain's summit. Reanalyses make use of modern numerical weather prediction systems to assimilate historical observations into consistent and homogeneous datasets suitable for diagnostic and model validation studies (Kalnay et al., 1996; Uppala et al., 2005). Comparisons with observations at 5800 m and 8000 m in the vicinity of Mount Everest indicate that reanalyses are able to capture the day-to-day and seasonal pressure and temperature variability in the region (Moore and Semple, 2004; Xie et al., 2009). Unfortunately, there are no validation studies with respect to wind speed, and it has been suggested that wind speeds in reanalyses near the summit of Mount Everest may be underestimated by a factor of 2 (Moore and Semple, 2011). The ERAI data is available four times daily for the period of 1979–2010.

These meteorological time series and expressions for WCT and FFT (Moore and Semple, 2011; Office of the Federal Coordinator for Meteorological Research, 2003) were used to calculate the hypothermic risk parameters during May 2006. Estimates of the climatological mean values for these parameters, as well as daily variability in them, were also calculated using the full 32 years of the dataset. The measure of daily variability is the standard deviation across the 32 years of values for a given time.

Results

In Figure 1 we show time series of the summit barometric pressure, temperature, and wind speed during May 2006 as extracted from the ERAI. Also shown are climatological mean values and an estimate of the daily variability in these parameters. As discussed previously, there is a tendency for the pressure and temperature to increase and for the wind speed to decrease during the spring as the weather transitions from a regime characterized by the influence of extra-tropical storms and the subpolar jet stream to one influenced by the Indian Summer Monsoon (Moore and Semple 2004, 2006; Moore et al., 2011). The night of May 14/15, Sharp's bivouac had a minimum in summit barometric pressure and temperature and a maximum in wind speed. Only the summit temperature had a departure from the mean that exceeded one standard deviation. In contrast during Hall's bivouac, the night of May 25/26, the summit barometric pressure and temperature were above the mean, while the summit wind speed was a minimum. All these parameters had departures from the mean that exceeded one standard deviation.

FIG. 1.

Time series of: (A) summit pressure (mb); (B) summit temperature (^oC), and (C) summit windspeed (m s⁻¹) for May 2006. Also shown by the dotted lines are the mean values, as well as one standard deviation above and below. The crosses indicate midnight local time on May 15 and 26.

In Figure 2, we show the associated hypothermic risk parameters during May 2006. As discussed previously (Moore and Semple, 2011), WCT and FFT are increasing during the spring as a result of the tendency towards warmer temperatures and lower wind speeds. The WCTD is decreasing over this period as the difference between the actual temperature and WCT decreases. During the night of May 14/15, the WCT was at ∼–51°C; while the FFT was ∼ 5 min. Both were below climatological mean values for this date with the WCT being lower by ∼1 standard deviation. The WCTD was ∼21°C. In contrast during the night of May 25/26, the WCT and FFT were ∼−31°C and 15 min, respectively. Both were in excess of 1 standard deviation above the climatological mean for this date. The WCTD was only ∼−8°C, well in excess of 1 standard deviation below the climatological mean for this date.

FIG. 2.

Time series of: (A) summit wind chill temperature (^oC); (B) summit wind chill temperature deficit (^oC), and (C) summit facial frostbite time (min) for May 2006. Also shown by the dotted lines are the mean values, as well as one standard deviation above and below. The crosses indicate midnight local time on May 15 and 26.

Discussion

We have presented time series of the environmental and hypothermic risk parameters for the summit region of Mount Everest during May 2006. In this month there were two high profile events, on May 14/15 and May 25/26, during which a solo climber was forced to bivouac above 8500 m. In the earlier event, the climber died, while in the later event, the climber survived.

Our results indicate that the two climbers experienced very different hypothermic environments that were the result of two distinct environmental processes. First of all, the 10-day difference between the two events resulted in higher climatological mean temperatures and lower climatological mean wind speeds that resulted in a reduced risk of hypothermia. In addition, the environmental conditions during the night of May 14/15 were considerably more severe than those during the night of May 25/26. This resulted in a considerably higher risk of hypothermia during the earlier event. Indeed, the differences in WCT, WCTD, and FFT between the two events were on the order of 2 standard deviations.

It is important to note that neither of these events occurred during stormy conditions and therefore it is not only during such extreme conditions that one must be concerned about the risks of hypothermia on Mount Everest.

These results suggest that the survival and death in these two similar cases was in a large part the result of very different hypothermic environments. Caution must however be exercised as the sample size in this study is small and there may have been other physiological factors, including hypoxia, dehydration, fatigue, and nutritional status, which could have influences the outcomes (Ainslie and Reilly, 2003; Ward, 1974). In addition, there are other variables, such as clothing and degree of exposure to the win, which also could have had an impact. It is therefore clear that there exists the need to extend these present findings with a larger sample.

In addition, the May 14/15 event occurred when the summit barometric pressure was low, while the May 25/26 event occurred when the summit barometric pressure was high. This supports the proposal that summit barometric pressure can act as a simple and easy-to-observe predictor for hypothermia and frostbite (Moore and Semple, 2011). In addition, to its use as a predictor for hypothermia and frostbite, the higher barometric pressure during the May 25/26 event would have also resulted in a reduction in hypoxic stress.

It is now routine for climbers on Mount Everest to make use of weather forecasts to provide guidance as to favorable summit conditions. These forecasts also provide all the information needed to provide guidance as to risk of hypothermia as well. We recommend that such guidance be made available to climbers so that they may have a better idea as to the risks that they are subject to so that they may plan accordingly. On the mountain itself, barometric pressure, which is a simple, easy to observe, and robust meteorological parameter, can be a useful predictor as to the risk of hypothermia that could be provided through tables that encompass the correlation between barometric pressure and WCT and FFT.

Footnotes

Acknowledgments

The authors would like to thank the European Center for Medium Range Weather Forecasting for kindly providing access to their Interim Reanalysis data set and to the reviewers for their comments.

Author Disclosure Statement

Dr. Moore was supported by the Natural Science and Engineering Research Council of Canada. Semple has no conflicts of interest or financial ties to report.

References

Ahmad

, Boraas

. 2010. Flow and thermal errors in current national weather service wind chill models. Heat Transfer Eng, 31:1137–1154.

Ainslie

, Reilly

. 2003. Physiology of accidental hypothermia in the mountains: A forgotten story. Br J Sports Med, 37:548–550.

Bjarnason

, Jardine

. 2005. 'I've never seen a storm hit like that': Climber to lose fingers to frostbite. National Post, 7 May 30, 2005 Toronto, Ontario, Canada.

D'Oro

. 2005. Blizzard traps Canadian climbers. Anchorage Daily News, 1 May 29, 2005 Anchorage, Alaska.

Dee

et al. 2011. The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart J Royal Meteorol Soc, 137:553–597.

Firth

et al. 2008. Mortality on Mount Everest, 1921–2006: Descriptive study. Br Med J, 337:a2654.

Hall

. 2008. Dead Lucky: Life after Death on Mount Everest. Jeremy P. Tarcher/Penguinxx, 30916of plates pp.

Harirchi

et al. 2005. Frostbite: Incidence and predisposing factors in mountaineers. Br J Sports Med, 39:898–901.

Heil

. 2008. Dark Summit: The True Story of Everest's Most Controversial Season. 1st Henry Holtxi, 2718p. of plates pp.

10.

Huey

, Eguskitza

. 2001. Limits to human performance: Elevated risks on high mountains. J Exp Biol, 204:3115–3119.

11.

Kalnay

et al. 1996. The NCEP/NCAR 40-year reanalysis project. Bull Am Meteorol Soc, 77:437–471.

12.

Kodas

. 2008. High Crimes: The Fate of Everest in an Age of Greed. 1st Hyperion, 357.

13.

McIntosh

et al. 2008. Mountaineering fatalities on Denali. High Alt Med Biol, 9:89–95.

14.

Moore

GWK

, Holdsworth

. 2007. The 25-27 May 2005 Mount Logan storm. Part I: Observations and synoptic overview. J Hydrometeorol, 8:590–606.

15.

Moore

GWK

, Semple

. 2004. High Himalayan meteorology: Weather at the South Col of Mount Everest. Geophys Res Lett, 31:18.

16.

Moore

GWK

, Semple

. 2006. Weather and death on Mount Everest. An analysis of the “Into Thin Air” storm. Bull Am Meteorol Soc, 87:465–480.

17.

Moore

GWK

, Semple

. 2011. Freezing and frostbite on Mount Everest: New insights into wind chill and freezing times at extreme altitude. High Alt Med Biol, 12:271–275.

18.

Moore

GWK

et al. 2010. Mallory and Irvine on Mount Everest: Did extreme weather play a role in their disappearance? Weather, 65:215–218.

19.

Moore

GWK

et al. 2011. Global Warming, EI Nino, and high-impact storms at extreme altitude: Historical trends and consequences for mountaineers, J. Appl Meteorol Climatol, 50:2197–2209.

20.

Osczevski

, Bluestein

. 2005. The new wind chill equivalent temperature chart. Bull Am Meteorol Soc, 86:1453.

21.

Office of the Federal Coordinator for Meteorological Research. 2003. Report of Wind Chill Temperature and Extreme Heat Indicies: Evaluation and Improvement Projects, 75.

22.

Pandolf

, Burr

. 2002. Medical aspects of harsh environments. Office of the Surgeon General United States Army.

23.

Pascual

, Callado

. 2010. Mountain accidents associated with winter northern flows in the Mediterranean Pyrenees. Tethys, 7:41–55.

24.

Siple

, Passel

. 1945. Measurements of dry atmospheric cooling in subfreezing temperatures. Proc Am Philos Soc, 89:177–199.

25.

Steadman

. 1979. The assessment of sultriness. Part II: Effects of wind, extra radiation and barometric pressure on apparent temperature. J Appl Meteorol, 187874–885.

26.

Subedi

et al. 2010. Frostbite in a Sherpa. Wilderness Environ Med, 21:27–29.

27.

Tikuisis

, Osczevski

. 2002. Dynamic model of facial cooling. J Appl Meteorol, 41:1241–1246.

28.

Uppala

et al. 2005. The ERA-40 re-analysis. Quart J Royal Meteorol Soc, 131:2961–3012.

29.

Ward

. 1974. Frostbite. Br Med J, 1:67–70.

30.

Xie

et al. 2009. Feasibility comparison of reanalysis data from NCEP-I and NCEP-II in the Himalayas. J Mountain Sci, 6:56–65.