WEATHER and DEATH on MOUNT EVEREST an Analysis of the Into Thin Air Storm

WEATHER AND DEATH ON MOUNT EVEREST An Analysis of the Into Thin Air Storm

BY G. W. K. MOORE AND JJOH N L. SEMPLE

A reconstruction of the meteorological condition leading to the deadly 1996 storm suggests promise for improving the predictability of mountain weather hazards and raises caution about the dangers of stratospheric ozone intrusions.

ith a height of8,848 m, Mount Everest is the highest mountain in the world. For over a century, it has been the subject of Wexploration for both scientific and recreational purposes ABOVE. Clouds approach the (Venables 2003). This exploration began in earnest with the British ex- Tibetan (north) side of Mount peditions during the early part of the twentieth century. At that time, Everest in the evening sun, Nepal was closed to foreigners and so these efforts were focussed on signaling a possible change in finding a route to the summit from the north through Tibet (Hemmleb the weather for the following et al. 2001; Venables 2003). During the 1921 expedition, a potential day. Storms on Everest are a route to the summit was identified that is now referred to as the North significant risk for mountaineers, appearing without warning and Col-North Ridge route (Fig. 1). It was via this route during the 1924 causing falling temperatures, expedition that Norton attained an altitude of 8,565 m without the increased winds, and heavy use of supplemental oxygen—a record that stood for 54 years until snowfall. [Photograph by J. L. Messner and Habler reached the summit in 1978 (Messner 1999; West Semple, May 2005.] 2000). Several days after Norton's feat, •

AMERICAN METEOROLOGICAL SOCIETY Unauthenticated | DownloadedAPRIL 200 09/24/216 BAH5 01:25- | 46 PM5 UTC this route until the beginning of World War II. After the war, as a result of turmoil in Tibet and the opening of Nepal to foreigners, the focus of the exploration of Mount Everest shifted to the south and the discovery of a route that followed the large horseshoe- shaped cirque or valley known as the Western Cwm, bounded by Nuptse, Lhotse, and Everest to the South Col, and then ultimately to the summit (Venables 2003). This route, known as the South Col route (Fig. 1), as- cends through the Khumbu Glacier and Kumbu Ice Fall into the Western Cwm, up and across the Lhotse Face to the South Col, and then along the south ridge to the summit of Everest. Please refer to Fig. 2 for a three- dimensional representation of Mount Everest developed by Gruen and Murai (2002) that provides further information on this route. This route FIG. I. Image of Mount Everest taken from the space shuttle I Oct was the one used by Hillary and 1993. In this striking early morning image, the north and south faces of Mount Everest as well as the Western Cwm, the large cirque to the Norgay (Hillary 2003) in the first west of Mount Everest, are in shadows while the Kangshung, or eastern, successful summit of Mount Everest face, is in bright sunlight. The South Col is situated between the sum- in May 1953. Although other routes mit of Everest and Lhotse, while the North Col is situated between the to the summit have been pioneered summit and Bei (or Changtse). Typical climbing routes via the South in the intervening years, these two and North Col are indicated. [NASA photograph STS058-I0I-I2.] routes remain the most popular (Huey and Salisbury 2003). Mallory and Irvine perished during a summit at- Much of the focus of scientific interest in Mount tempt made with the use of supplemental oxygen Everest has been on the impact that the low barometric (Hemmleb et al. 2001; Venables 2003). To this day, pressure near its summit has on human physiology it is still unclear if Mallory and Irvine were able to (Houston et al. 1987; West 1999; Huey and Eguskitza reach the summit before their deaths (Hemmleb 2001). These studies have shown that above 7,000 m, et al. 2001; Messner and Carruthers 2001; Hemmleb climbers are at the limits of their endurance and are et al. 2002). Exploration of Everest continued via exposed to significant risks associated with the reduced amount of oxygen available for respiration: cold temperatures and high winds (West et al. 1983a; West 2000; AFFILIATIONS: MOORE—Department of Physics, University Huey and Eguskitza 2001; Huey et al. 2001). of Toronto, Toronto, Ontario, Canada; SEMPLE—Department of Perhaps the most dramatic example of these risks Surgery, University of Toronto, Toronto, Ontario, Canada CORRESPONDING AUTHOR: Prof. G. W. K. Moore, occurred in early May 1996 when a storm engulfed Department of Physics, University of Toronto, 60 George Street, Mount Everest, trapping a number of climbers on Toronto ON M5S IA7, Canada. its exposed upper slopes and ultimately resulting E-mail: [email protected] in eight deaths.1 This event and the ensuing tragedy

The abstract for this article can be found in this issue, following the 1 table of contents. It has become conventional to refer to this event as a "storm," DOI: 10.1175/BAMS-87-4-465 although as we shall show it was in all likelihood an extended

In final form 11 October 2005 outbreak of unfavorable weather with two distinct periods ©2006 American Meteorological Society of very stormy weather. For simplicity, we will refer to the event as the Into Thin Air storm.

466 I BAflS- APRIL 2006 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC received international attention at the time (Burns 1996; Horsnell and Faux 1996) and became the subject of a number of books, including Into Thin Air (Krakauer 1999) and others (Boukreev and DeWalt 1999; Breashears 1999; Dickinson 1999), as well as figuring prominently in the movie "Everest" (MacGillvray and Breashears 1998). On the evening of 9 May 1996,2 a large number of climbers were poised to make summit attempts from camp IV situated at 8,000 m on the South Col of Mount Everest (Fig. 2). High winds had persisted throughout the day and the possibility of summitting appeared low. The winds died down during the evening and the conventional wisdom was that they would remain calm for a period of time. As a result, FIG. 2. Three-dimensional photorealistic visualization of Mount the decision was made to attempt to Everest showing features and approximate elevations along the South summit and the climbers left around Col route. Visualization courtesy of A. Gruen, Institute of Geodesy 2000 LT on the 9th for the 18-24-h and Photogrammetry, Swiss Federal Institute of Technology. round-trip to the summit. During the afternoon of 10 May, an intense storm3 with wind that had taken shelter were suffering from the lack speeds estimated to be in excess of 30 m s_1, heavy of oxygen that, as will be discussed in this paper, snowfall, and falling temperatures engulfed Mount may have been exacerbated by the unusually low Everest, trapping over 20 climbers on the exposed barometric pressure. A number of climbing parties upper sections of the mountain above the South Col who were attempting to summit along the North (Fig. 2). The unfavorable weather appears to have Col-North Ridge route (Fig. 1) were also trapped by abated overnight, although winds remained high near the storm and three people died. Although the harsh the summit. There was however a reintensification weather abated after the 11th, winds near the summit during the day on the 11th. Throughout this period, remained high for next 5 days, hindering attempts a number of heroic attempts were undertaken to res- to summit. cue the trapped climbers that were hampered by the In Fig. 3, infrared imagery from the National high winds and the harsh weather. Tragically, five of Oceanic and Atmospheric Administration (NOAA) the climbers could not be rescued and perished. By family of polar-orbiting satellites are presented for the afternoon of the 11th, most of the people on the the late afternoon during the period of 9-12 May upper slopes of Mount Everest were sheltered in tents 1996. The images are derived from the Global Area at the South Col (Fig. 2). In this regard, the impact Coverage (GAC) level-IB dataset (Kidwell 1998) and of the weather on the 11th was not as catastrophic as have a resolution of 4 km. Brightness temperatures that on the 10th. Supplemental oxygen was however derived from the observed radiances are presented in short supply and there is evidence that even those with a color map in which colder (and therefore

2 The following narrative is based on the published recollections of the storm as contained in Boukreev and DeWalt (1999), Dickinson (1999), and Krakauer (1999). All times are given in local Nepalese time, which is 5 h and 45 min ahead of coordi- nated universal time. 3 Among those on the upper slopes of Everest was a pilot who was familiar with the appearance of thunderstorms as seen from above. He is quoted in Krakauer (1999) as stating that he believed that the clouds that engulfed Mount Everest were associated with the crown of a "robust" thunderstorm. In two separate incidents during the storm, Krakauer (1999) reports that he and a sherpa observed thunder and lightning.

AMERICAN METEOROLOGICAL SOCIETY APRIL 2006 BAF15- | 467 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC activity in the vicinity of Mount Everest. This outbreak of high- impact weather resulted in the largest number of fatalities (8) to occur near the summit of Everest during a single event. In the ensuing years, a passionate and wide-ranging debate has been ongoing in an attempt to understand the factors that contributed to this tragedy (Boukreev and DeWalt 1999; Breashears 1999; Dickinson 1999; Krakauer 1999; Roberto 2002; Mangione 2003; Kayes 2004). Curiously absent has been a quan- titative discussion of the meteorological conditions that gave rise to this storm and the storm's impact on the climbers' physiology. Motivated by the 1996 tragedy on Mount Everest, the Media Laboratory at FIG. 3. Brightness temperatures (°C) derived from NOAA-12 infrared satellite images from (a) 1242 UTC (1827 LT) 9 May 1996, (b) 1221 UTC (1806 LT) the Massachusetts Institute 10 May 1996, (c) 1159 UTC (1744 LT) II May 1996, and (d) 1317 UTC (1922 LT) of Technology (MIT) de- 12 May 1996. The location of Mount Everest is indicated by the +. veloped a portable sen- sor package to monitor weather conditions on higher) objects appear brighter. The images from Mount Everest (Lau 1998). One such portable weath- 9 May 1996 (Fig. 3a) show widespread and scattered er station was deployed at the South Col of Mount convection over the Tibetan Plateau with India being, Everest and operated from May to September 1998. for the most part, cloud free. The distinct boundary Moore and Semple (2004) performed an analysis of between the cloud-covered and cloud-free areas lies these data and compared the observed temperatures along the Himalayas (Hirose and Nakamura 2005). and pressures to both the National Centers for Envi- The images from 10 and 11 May (Figs. 3b and 3c), the ronmental Prediction (NCEP)-National Center for period of the Into Thin Air storm, are dramatically Atmospheric Research (NCAR) reanalysis (Kalnay different with the presence of deep convection in the et al. 1996) and the European Centre for Medium- immediate vicinity of Mount Everest. The outbreak of Range Forecasts (ECMWF) Re-Analyses (ERA-40) convection on the 11th appears more organized, with (Simmons and Gibson 2000). They found that the the clouds tops higher than those on the 10th. Indeed, reanalyses were able to capture much of the day- the convection in the vicinity of Mount Everest on to-day variability in the pressure and, to a lesser the 11th has the characteristic circular cloud shield extent, the temperature at the South Col of Mount or anvil that is often associated with organized col- Everest. A high-impact weather event in early May, lections of thunderstorms known as mesoscale con- which resulted in a major impact to operations on vective complexes (Laing and Fritsch 1993). On the Mount Everest with high winds and heavy snowfall 12th (Fig. 3d), one has a situation similar to the 9th, reported (W. Berg 1998, personal communication), with widespread but scattered and shallow convection was captured in the South Col meteorological over the plateau with no evidence of deep convective observations. Moore and Semple (2004) showed

468 I BAflS- APRIL 2006 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC that this event was associated with the passage of an upper-tropospheric jet streak, an elongated region of high wind speed embed- ded within the subtropical jet stream, near Mount Everest. Moore and Semple (2004) also showed that this event from 1998 was associated with a maximum in the time series of the Total Ozone Map- ping Spectrophotometer (TOMS) total column ozone over Mount Everest and furthermore suggest- ed that ozone-rich stratospheric air may have been present near the summit during this event. These results suggest that the 1 FIG. 4. Variability in the (a) temperature (°C), (b) wind speed (m s ), (c) pres- summit of Mount Everest sure (mb), and (d) potential vorticity (PVU) at the summit of Mount Everest resides in the dynamically during May as extracted from the ERA-40 reanalysis. The solid black lines interesting but relatively indicate the mean values over the period of 1958-2002. The dashed black lines unexplored region of the delimit the band that contains 90% of all of these values during the period atmosphere known as the of 1958-2002. Therefore, values above/below the upper/lower black dashed lines occur in approximately 5% of the years. The red lines indicate the values "middleworld" (Hoskins during May 1996. These values are shown every 6 h with the * indicating the 1991), where the strato- data at 0000 UTC each day. sphere and troposphere are in direct communication. During the 1998 high-impact weather event, circulation patterns. In this paper, we will draw upon there was a relatively large discrepancy between these results to investigate the weather associated the observed pressure at the South Col and the ex- with the May 1996 Into Thin Air tragedy and its im- tracted values from the NCEP-NCAR and ERA-40 pact on the people who were trapped near the summit reanalyses. Through the analysis of satellite imagery of Mount Everest. of the event, Moore and Semple (2004) argued that this discrepancy was the result of an outbreak of A SYNOPTIC-SCALE ANALYSIS OF THE convection in the vicinity of Mount Everest that INTO THIN AIR STORM. We begin our analysis was not resolved in the reanalyses. With regard to with Fig. 4 in which the time series for temperature, the initiation of this convection, they that argued it wind speed, pressure, and potential vorticity at the was triggered by ageostrophic circulation associated summit of Mount Everest during May 1996 were ex- with the presence of the jet streak in the vicinity of tracted from the ERA-40 reanalysis. All time series Mount Everest (Uccellini and Johnson 1979; Keyser were obtained by interpolation of the corresponding and Shapiro 1986). three-dimensional fields to the location and height From Moore and Semple's (2004) analysis of of Mount Everest (27°59'N, 86°55'E; 8,850 m). Also the South Col observations and other work (Lang shown are climatological mean time series and es- and Barros 2004; Moore 2004; Kennett and Toumi timates of the interannual variability based on the 2005), it is clear that the NCEP-NCAR and ERA-40 time period of the ERA-40 reanalysis (1958-2002). reanalyses are useful datasets with which to study the The first three fields are obvious ones that describe meteorology of the Himalayas. In addition, it appears the general meteorological conditions at the summit. that at least some of the high-impact weather that Potential vorticity is in a different class because it is, affects Mount Everest is associated with large-scale for example, not directly observable. It is however a

AMERICAN METEOROLOGICAL SOCIETY APRIL 2006 BAF15- | 469 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC relatively simple and widely used diagnostic that al- decrease in wind speed, an increase in pressure, lows one to distinguish between air of tropospheric and a decrease in potential vorticity occured dur- and stratospheric origins (Danielsen 1984; Hoskins ing the month of May. As described by Moore and 1991; Morgan and Nielsen-Gammon 1998), with Semple (2004), these changes are associated with a values in the range of 1-2 potential vorticity units warming of the troposphere and the transition to (1 PVU = 10-6 K s 2 kg-1) usually serving as the cut- the oncoming Indian summer monsoon. All fields off. However, as discussed by Hoskins (1991), the are characterized by a decrease in variability during exact value used as a cut-off is not as important as May. With regard to the temperature and pressure, the jump in potential vorticity that exists across the the outliers contain the same basic trend as does tropopause. the mean. In contrast, the trend in variability of As discussed in Moore and Semple (2004), there the wind speed and potential vorticity appears to will be variability in these fields that is not resolved be the result of a reduction in the occurrence of in the ERA-40 reanalysis as a result of its coarse anomalously high wind speed/potential vorticity resolution. These time series should therefore be events. We believe this to be also a consequence of interpreted as being representative of the synop- the oncoming monsoon that results in a transition tic-scale variability in these fields. With regard from a regime impacted by the passage of baroclinic to the climatology, we observe that a warming, a weather systems to a more convective regime. Turning our attention to the events in May 1996, bearing in mind that the ERA-40 reanalysis captures only the synoptic-scale variability, we see that the summit temperature underwent an approximate 10°C drop from the 8th to the 13th. This transition was associated with a change from anomalously warm to anomalously cold temperatures. Temperatures remained low until approximately 25 May when they returned to values seen in early May. The summit wind speed time series indicates the presence of several distinct maxima, followed by minima that occur with a period of approximately 1 week. One of these maxima occurred on the 9th when wind speeds at the summit were approximately 22 m s_1. The wind speeds decreased over the next 2

FIG. 5. Spatial structure of the climatological mean (a) geopotential height (contours, km), potential vorticity (shading, PVU), and (b) potential vorticity (contours, PVU) and horizontal wind (vectors, m s-1) and wind speed (shading, m s1) fields on the 340-K potential temperature surface during May. Climatology based on the ERA-40 reanalysis for the period of 1958-2002. The 4,000-m height contour is indicated with the dashed line, while the + indicates the location of Mount Everest.

470 I BAflS- APRIL 2006 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC days to approximately 15 m s"1 before increasing to values in excess of 25 m s_1 on the 12th. Wind speeds on the 12th exceeded those observed in 95% of the years in the reanalysis record. The summit pressure was anomalously high in early May and then decreased over the next 8 days, reaching anomalously low values on 12 and 13 May. The pressure remained low for the next 10 days before returning to anomalously high values around 25 May. With regard to potential vorticity, we see that a transition from tropospheric, that is, < 1 PVU, to stratospheric, that is, > 1 PVU, values occurred between 8 and 10 May. The potential vorticity remained high for approximately FIG. 6. Spatial structure of the climatological mean vertically integrated moisture transport field (vectors, kg m_l s_l) and its magnitude 3 days before returning to lower (contours and shading, kg nrr1 s_l) during May. Climatology based values. There were two additional on the ERA-40 reanalysis for the period of 1958-2002. The 4,000-m events later in May when high values height contour is indicated with the dashed line, while the + indicates of potential vorticity occurred. the location of Mount Everest. We now consider the spatial representation of the synoptic-scale flow during In Fig. 6, we present the climatological mean verti- the Into Thin Air storm. Motion in the upper tro- cally integrated moisture transport field and its mag- posphere, where the summit of Mount Everest is nitude during May. This field, defined as the vertical located, is quasi adiabatic and occurs, to first order, integral of the product of the specific humidity and along surfaces of constant potential temperature the horizontal wind, provides a succinct view of the (Hoskins et al. 1985). Therefore, to visualize the atmospheric water balance, with convergence of the large-scale circulation during this event, we have field on monthly and longer time scales being propor- chosen to view the flow on the 340-K potential tional to the excess of precipitation over evaporation temperature surface. (Peixoto and Oort 1992; Smirnov and Moore 1999). Figure 5 shows the climatology on this surface This figure shows that during May there are two during May for the period of 1958-2002. With regard main moisture streams in the region—the dominant to the climatological mean height of this surface stream that transports moisture in a northeasterly (Fig. 5a), we see that Mount Everest is situated in the direction over the Bay of Bengal toward southeast region of large meridional gradient as one moves Asia, and a weaker stream that transports moisture northward from the Tropics toward the midlatitudes. eastward from the Arabian Sea over northern India The height of this surface in the vicinity of Mount south of the Himalayas. There appears to be some Everest during May is approximately 9.25 km. From convergence effects associated with the high topog- the climatological mean potential vorticity on this raphy of Tibet and the merging of the two streams surface (Figs. 5a, and 5b), it is clear that the summit of over northeast India and Bangladesh, which receives Mount Everest is typically in the troposphere with the the heaviest rainfall on the subcontinent during May tropopause just to the north over central Tibet. This (Parthasarathy et al. 1994). confirms that Mount Everest lies in the middleworld We present in Figs. 7 and 8 the spatial distribution (Hoskins 1991). From the climatological mean wind of the potential vorticity, height, and wind speed in speeds on this surface (Fig. 5b), one can see that the the 340-K potential temperature surface before, dur- region of high wind speeds known as the subtropi- ing, and after the event. At 0000 UTC 7 May (Figs. 7a cal jet stream is typically situated just to the north of and 8a), the height of this surface in the vicinity of Mount Everest during May. The region of low wind Mount Everest was approximately 8.5 km, which is speeds in this jet stream over Tibet was first noted by lower than the climatological value of approximately Krishnamurti (1961). 9.25 km (Fig. 5a). This reflects the higher pressures that

AMERICAN METEOROLOGICAL SOCIETY APRIL 2006 BAF15- | 471 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC were occurring near the summit (Fig. 4c). At this time, During the storm, 0000 UTC (0545 LT) 11 May Mount Everest was situated in the troposphere with (Figs. 7b and 8b), the height of the 340-K surface in potential vorticities in its vicinity less than 1 PVU. The the vicinity of Mount Everest was approximately wind speed field (Fig. 8a) clearly shows the presence of 10 km, reflecting the lower pressures that were pres- a jet streak just to the north of Mount Everest with its ent near the summit at this time (Fig. 4c). This lower entrance region to the west of Mount Everest. pressure was associated with an upper-level shortwave trough that was situated just to the west of Mount Everest at this time (Schultz and Doswell 1999). The source of the elevated high potential vorticity identified in Fig. 4d can be seen from Figs. 7b and 8b to be a filamentary intrusion of high potential vorticity air that extended from the midlatitudes toward the Mount Everest region. Such intrusions are favored in northwesterly flow behind shortwave troughs (Rotunno et al. 1994; Schultz and Doswell 1999). Evidence of this intrusion and the associated shortwave trough can be seen, albeit in a weaker form, over the Arabian Sea on the 7th (Figs. 7a and 8a). Two jet streaks were in the vicinity of Mount Everest on 11 May (Fig. 8b). In addition to the jet streak present on 7 May (Fig. 7a), another jet streak was situated to the west of Mount Everest on the 11th. As a result of its eastward propagation, the entrance region of the jet streak identified in Fig. 7a was now situated to the northeast of Mount Everest, while the exit region of the second jet streak was to the southwest. Over the next several days, the shortwave trough and the filamentary intrusion of high potential air moved to the east and continued to intensify into a well-defined tropopause fold by the 15th (Figs. 7c and 8c). On this date, Mount Everest was again situated in the troposphere

FIG. 7. Spatial structure of the geopotential height (contours, km), the potential vorticity (shading, PVU), and the horizontal wind (vectors, s_l) fields on the 340-K potential temperature surface at 0000 UTC on (a) 7 May, (b) 11 May, and (c) 15 May 1996. All fields are from the ERA-40 reanalysis. The + indicates the location of Mount Everest.

| BAH* APRIL 2006 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC with the 340-K surface being at a height of 9.5 km, entrance region (Keyser and Shapiro 1986). Convec- close to the climatological value of 9.25 km. The sec- tive activity, which is the result of an ageostrophic ond jet streak that was situated to the west of Mount circulation that develops in the plane perpendicular Everest on the 11th was now situated to its north, to the jet streak, preferentially occurs to the left of the resulting in the high winds that we observed near the exit region and to the right of the entrance region in summit in the aftermath of the storm (Fig. 4b). a coordinate system in which the observer is looking The synoptic situation presented in Figs. 7 and 8 provides an explana- tion for the observations of the people who witnessed strong winds on 9 May, (associated with the eastern jet streak) and a weakening of these winds late on the 9th as this streak propagated away from Everest, and then a reintensification of the high winds a few days later as the second jet streak passed by. The recollections of those near the summit during the event clearly indicate the presence of high winds in addition to other severe weather elements. This is inconsistent with the picture that we have presented of the synoptic-scale conditions during the storm. Satellite imagery during the event (Figs. 3b and 3c) clearly shows the outbreak of deep convection during the event. In the following section, we show that this convective component to the event was forced by the synoptic-scale flow in the region.

CONVECTION IN THE VICIN- ITY OF MOUNT EVEREST DURING THE INTO THIN AIR STORM. Convection is often observed to occur in conjunction with jet streaks (Uccellini and Johnson 1979; Keyser and Shapiro 1986; Doswell and Bosart 2001). The section of a jet streak where the air parcels are decel- erating is referred to as the exit region, while the section where the air parcels are accelerating is referred to as the

FIG. 8. Spatial structure of the potential vorticity (contours, PVU) and horizontal wind (vectors, m s_l) and wind speed (shading, m s~') fields on the 340-K potential temperature surface at 0000 UTC on (a) 7 May, (b) 11 May, and (c) 15 May 1996. All fields are from the ERA-40 reanalysis. The + indicates the location of Mount Everest.

AMERICAN METEOROLOGICAL SOCIETY APRIL 2006 BAF15- | 67 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC in the direction that the wind is blowing (Uccellini ERA-40 reanalysis. At that time, there is evidence and Johnson 1979; Keyser and Shapiro 1986). Indeed, of two distinct circulations emanating from a re- the satellite imagery (Figs. 3b and 3c) clearly shows gion of divergence that was situated just to the east that an outbreak of deep convection near Mount of Mount Everest. One circulation had a region of Everest occurred during this event. In this section, we convergence to the west of Mount Everest, while the will argue that this convection was the result of the other has a region of convergence to the northeast. juxtaposition of the two jet streaks, enhanced by the The positions of the jet streaks at this time (Fig. 8b) presence of the shortwave trough and the large flux of show that the former straddles the entrance region of water vapor in the vicinity of Mount Everest. We will the jet streak to southwest, while the latter straddles focus our attention on 0000 UTC (0545 LT in Nepal) the exit region of the jet streak to the northeast. The 11 May, a time that represents the approximate middle effect of curvature associated with the presence of a point in the outbreak of harsh weather. shortwave trough is to enhance divergence to the east To verify the presence of ageostrophic circulation, of the trough and divergence to the west (Schultz and we use the Helmholtz theorem (Arken 1985) to parti- Doswell 1999), and thus would have acted to enhance tion the horizontal wind field into rotational and the ageostrophic circulation associated with the ap- V V r divergent Vd components, proaching jet streak. Advection of moisture into the region where there

V=V + Vd = kxV y+Yfr (1) is forced ascent is also a requirement for an outbreak of severe convection, preferably in a direction per- where f is the streamfunction, x is the velocity po- pendicular to the axes of the jet streaks (Uccellini and tential, and A: is a unit vector in the vertical direction. Johnson 1979). In Fig. 10, we present the vertically in- In this formulation, the jet streak-induced motion tegrated moisture transport and its magnitude as well

is contained in the divergent wind field Vd with as the total precipitation at 0000 UTC 11 May 1996 as maxima/minima in the velocity potential associated depicted in the ERA-40 reanalysis. It should be noted with regions of convergence/divergence. This parti- that the precipitation in the ERA-40 reanalysis is rela- tion was performed with the SPHEREPACK software tively unconstrained by observations and so is highly package (Adams and Swarztrauber 1999). dependent on the details of the underlying numerical In Fig. 9, we present the velocity potential % and weather prediction model, resulting in a tendency

the divergent wind field Vd on the 300-mb pressure for the ERA-40 precipitation be have higher values surface at 0000 UTC 11 May 1996 as depicted in the than those that are observed (May 2004; Andersson et al. 2005). Nevertheless, this bias does not appear to be a problem with respect to the Indian summer monsoon and Himalayan rainfall (Challinor et al. 2005; Kennett and Toumi 2005). On this date, the vertically integrated moisture transport field has a structure similar to the climatological mean for May (Fig. 6), but with a magnitude in the region of interest that is approximately twice as large. Examination of this field on earlier days confirms that the maximum to the southeast of Mount Everest is the result of the merging of two streams—one from the Bay of Bengal and the other from the Arabian Sea (not shown). There is also evidence of a FIG. 9. Spatial structure of the velocity potential (contours and shading, I06m2 s"1) and the horizontal divergent velocity (vectors, m s_l) channelling of the moisture flux along fields on the 300-mb pressure surface at 0000 UTC 11 May 1996. All the southern edge of the plateau. The fields are from the ERA-40 reanalysis. The + indicates the location total precipitation field has a maxi- of Mount Everest. mum of 18 mm day1 in the region;

474 I BAflS- APRIL 2006 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC the convective component of this precipitation was relatively small at less than 4 mm da)^1 (not shown). The magnitudes of the vertically integrated moisture flux and total precipitation in the region to the southeast of Mount Everest during this outbreak of harsh weather were exceeded only 5% of the time during May for the ERA-40 time period of 1958-2002. More information on the spatial structure of the atmospheric moisture stream to the southeast of Mount Everest is provided in Fig. 11. Here we show the zonal and meridional components of the water vapor transport field, defined as the product of the specific humidity and the horizontal velocity field (Peixoto and Oort 1992) along 92.5°E as depicted in the ERA- 40 reanalysis. The moisture in this stream is confined to the lower 4 km of the atmosphere with only a slight spillover onto the Tibetan Plateau, and with the northward transport being approximately 60% larger than the eastward transport. In Fig. 12, we show the zonal wind and vertical velocities, as well as the potential vorticity along a meridional cross section through Mount Everest (87°E) at 0000 UTC 11 May 1996, as depicted in the ERA-40 reanalysis. The high zonal wind speeds to the south of Mount Everest FIG. 10. Spatial structure of (a) the vertically integrated moisture were associated with the exit region transport field (vectors, kg m_l s-1) and its magnitude (contours and of the oncoming jet streak (Fig. 8b). shading, kg m_l s_l) and (b) the total precipitation (shading, mm The intrusion of high potential day1) fields at 0000 UTC May 1996. All fields are from the ERA-40 vorticity, that is, stratospheric, air reanalysis. The + indicates the location of Mount Everest. associated with this event can be seen to be associated with the presence of a tropo- The results presented in this section are consistent pause fold (Keyser and Shapiro 1986) in the vicinity with the hypothesis that the convection respon- of Mount Everest. With regard to the vertical velocity sible for the Into Thin Air storm was the result of field, there exists a dipole on this day with a region of synoptic-scale forcing. Finally, we note that there is rising motion in the vicinity of Mount Everest and a a pronounced diurnal cycle to the convection over region of descending motion over the Tibetan Plateau the Tibetan Plateau with the most intense activity to the north. This dipole is consistent with idealized occurring in the late afternoon and early evening representations of the transverse circulation that is (Laing and Fritsch 1993; Yanai and Li 1994). This is associated with the presence of the jet streak to the the time of day on both the 10th and 11th when the northeast of Mount Everest (Uccellini and Johnson high-impact weather on Mount Everest was most 1979; Keyser and Shapiro 1986; Doswell and Bosart intense (Boukreev and DeWalt 1999; Krakauer 1999). 2001), as well as with the divergent circulation il- This diurnal cycle may have also contributed to the lustrated in Fig. 9. high-impact weather.

AMERICAN METEOROLOGICAL SOCIETY APRIL 2006 BAF15- | 475 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC PHYSIOLOGICAL EFFECTS OF HIGH-IMPACT WEATHER ON MOUNT EVEREST. It has been well established that the hypoxia caused by the low barometric pressures at extreme altitudes may severely limit human performance. Observations from the summit of Mount Everest indicate an in-

spired partial pressure of 02 (P02) of 57 mb compared with a sea level value of 198 mb (West et al. 1983b). Furthermore, above 7,000 m extreme hyperventilation is required to pro-

vide even these low values of P02 (West et al. 1983b). This amount of oxygen is barely sufficient for the requirements to maintain even the basal metabolic rate of a human being (West 2000). This has lead to many physiological studies that have concentrated on the ability of acclimatized mountaineers to undergo extreme exertion while in this severe hypoxic state (West 1984; Houston et al. 1987; West 2000). One of the conclusions reached was that the available oxygen levels at these extreme altitudes were exquisitely sensitive to even small variations in barometric r .. ^ .. r/ w. . . J/LWL ... pressurer . Furthermore, in the physir 7 - FIG. 11. Cross section of (a) the zonal component and (b) the meridional of the moisture transport (shading and contours, g kg1 m s ') ological literature> the relationship along 92.5°E at 0000 II May 1996. All fields are from the ERA-40 between barometric pressure and al- reanalysis. Please note that the topography along the cross section titude has been assumed, apart from does not resolve the full height of topographic features. seasonal variations, to be relatively constant (West et al. 1983b; Hous- ton et al. 1991; West 1999; Huey and Eguskitza 2001; Huey et al. 2001). From the South Col observations of the May 1998 storm discussed by Moore and Semple (2004) as well as the results presented in this paper,

FIG. 12. Cross section of the vertical velocity (shading, cm s_l) and the zonal wind speed (solid contours, m s_l) and potential vorticity (dashed contours, PVU) though Mount Everest along 87°E at 0000 UTC II May 1996. All fields are from the ERA-40 reanalysis. Please note that the topography along the cross section does not resolve the full height of Mount Everest.

476 I BAflS- APRIL 2006 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC it is clear that there exist pressure fluctuations that storm. A similar conclusion was reached by Moore are associated the onset of high-impact weather in and Semple (2004). Ozone is recognized as posing a the Mount Everest region. From Fig. 4, one can see risk to human health through the impairment of lung that the Into Thin Air storm was associated with function that comes through the exposure to elevated an approximate 8-mb drop in the barometric pres- levels thereof (Devlin et al. 1996; Environmental sure at the summit of Mount Everest in the ERA-40 Protection Agency 1996). In this regard, it is interest- reanalysis. This drop in pressure was most likely as- ing to note that exposure to elevated ozone concentra- sociated with the presence of the shortwave trough. In tions at moderate altitudes (2434 m) in the Austrian a comparison between the ERA-40 and NCEP-NCAR Alps has been postulated to result in a decrease in pul- reanalyses and observations at the South Col during monary function in mountain guides (Wittels et al. the May 1998 storm, Moore and Semple (2004) found 1997). Furthermore, ozone of stratospheric origin as- an approximate 40% underestimation in the mag- sociated with tropopause folds has been measured in nitude of the pressure drop in the reanalyses. This commercial aircraft for over 40 yr at concentrations error was attributed to the fact that the reanalyses that can exceed the air quality guidelines (National did not capture the pressure fluctuations associated Research Council 2002; Spengler and Wilson 2003; with the convective component of the high-impact Spengler et al. 2004). Indeed, concentrations of ozone weather. A similar result is probably true for this in tropopause folds of the type that were associated event as well. The summit of Mount Everest is ap- with the Into Thin Air storm are in the range of 120- proximately 850 m above the South Col site. Based 280 ppb (Viezee et al. 1983; Davies and Schuepbach on the variation of pressure with height in both the 1994; Ravetta and Ancellet 2000; Roelofs et al. 2003). NCEP-NCAR and ERA-40 reanalyses, we estimate At these concentrations, which exceed guidelines that during the 1998 event, there was an additional established by the U.S. Environmental Protection 20%-30% increase in the magnitude of the anomaly Agency and other agencies, it seems reasonable to at the summit. Assuming similar conditions, we suggest that the presence of ozone near the summit estimate that climbers trapped above the South Col of Mount Everest represents a hitherto unrecognized would have experienced a total pressure drop between risk to climber's health. 10 and 14 mb during the Into Thin Air storm. Such a drop in barometric pressure would result CONCLUSIONS. In this paper, we have investi- in an approximate 6% reduction in inspired partial gated the meteorological conditions associated with P02 near the summit (West et al. 1983b; West 1999). the Into Thin Air storm of May 1996. This outbreak of Nonlinearities in the relationship between P02 and high-impact weather is by all accounts the deadliest maximum oxygen uptake (V02max) result in an ap- one to occur on Mount Everest, with eight deaths on proximate 14% reduction in the latter for climbers its upper slopes attributed to it. Our analysis builds near the summit (West 1999). To put these values on the results of Moore and Semple (2004) who used into context, similar drops in barometric pressure, observations collected at the South Col of Mount PO,, and VO, are estimated to occur between sum- Everest during 1998 to show that the global reanalyses 2 2max mer and winter near the summit of Mount Everest of NCEP-NCAR and the ECMWF can capture the (West et al. 1983b; West 1999). These reductions are synoptic-scale variability in the pressure and tem- of physiological significance, and according to some perature near the summit of Mount Everest. investigators are probably why Everest has never been Our results indicate that the high-impact weather climbed during midwinter without supplemental during the Into Thin Air storm was the result of an oxygen (West 1999). In view of these findings even outbreak of deep convection in the vicinity of Mount strong, experienced climbers who were caught near the Everest (Fig. 3). The outbreak occurred during a summit of Everest in a storm of this magnitude would transition from anomalously warm high pressure be at a physiological disadvantage. For those climbers to anomalously cold low pressure conditions during without supplemental oxygen or those who perhaps a period of low synoptic-scale wind speed that was were only partially acclimatized after running out of bracketed by periods of high wind speed (Fig. 4). In supplemental oxygen, the physiological impact of the addition, the event occurred during a period of high, barometric pressure drop associated with such a storm that is, stratospheric, values of potential vorticity at would have been most profound. the summit (Fig. 4). Detailed information on the In addition, we have presented evidence as to the meteorological conditions before, during, and after possible presence of ozone-rich stratospheric air on this event are unfortunately unavailable. However, the upper slopes of Everest during the Into Thin Air the narratives of Krakauer (1999) and Boukreev and

AMERICAN METEOROLOGICAL SOCIETY APRIL 2006 BAF15- | 477 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC DeWalt (1999) both mention the high winds that were circulations that could assist in determining the po- present near the summit on 9 May and the subsequent tential for harsh weather. weakening of the winds just prior to the outbreak There exists evidence that fatalities on Everest of the harsh weather. Dickenson (1999), who was typically occur during descent and that the prob- attempting to summit via the North Col-North Ridge ability of death is significantly higher for those not route and as a result was in a more exposed position using supplemental oxygen (Huey and Eguskitza vis-a-vis the westerly wind, comments on the sus- 2000; Huey et al. 2001). Operational constraints tained high winds that occurred after the outbreak, dictate that descents from the summit of Mount which rendered it impossible for him to summit until Everest typically occur in the late afternoon or early they weakened later in the month. evening (Breashears 1999; Krakauer 1999). As we We have also shown that there were two jet streaks have discussed, this is the period of the day when the and an upper-level shortwave trough present in the probability of convective activity is highest over the vicinity of Mount Everest during this event (Figs. 7 and adjoining plateau (Laing and Fritsch 1993; Yanai and 8). The positioning of these two streaks and the trough Li 1994). The higher probability of convective activity was such that the transverse ageostrophic circulation in the late afternoon when climbers are descending associated with them resulted in a synoptic-scale rising and near exhaustion is another risk factor that has motion near Mount Everest (Figs. 8 and 9). In addition, not been previously identified. there was in the lower troposphere an anomalously It is therefore likely that the falling pressure and large transport of water vapor from both the Arabian the high-impact weather that ensues may play a role Sea and the Bay of Bengal into the region to the south in the fatalities that occur on Mount Everest. This is of Mount Everest that was oriented approximately especially true above 7,000 m, where climbers are at perpendicular to the jet streaks (Figs. 10 and 11). We the limits of their endurance and drops in barometric believe that it was the juxtaposition of the ageostrophic pressure of the magnitude that we have established, circulation associated with these upper-level features when compounded by the accumulative effect of hy- near Mount Everest that in combination with the poxia, fatigue, high winds, extreme cold, and incom- anomalous availability of moisture triggered the orga- plete acclimatization, could shift a coping climber nized convective activity (Fig. 12). The harsh weather from a state of brittle tolerance to physiological occurred in the late afternoon, and it is likely that this distress. The presence of ozone-rich stratospheric air synoptically forced rising motion was enhanced by that near the summit of Mount Everest during the event associated with the diurnal cycle of solar heating. may have provided an additional stress on the climb- Our results, along with those of Moore and ers through a reduction in pulmonary function. Semple (2004), suggest that it may be possible to use the reanalysis datasets and the models upon which ACKNOWLEDGMENTS. Ian Renfrew and Mark they are based to study and forecast high-impact Stastna provided helpful comments on an earlier draft weather systems in the Mount Everest region, es- of the paper. We also thank the editor and reviewers for pecially in the premonsoon period. This is of great comments that greatly improved the paper. The ERA-40 practical benefit because this is the period during reanalysis data were provided by the European Centre for which much of the activity on the mountain takes Medium-Range Weather Forecasts. The satellite imagery place (Huey and Salisbury 2003). As an example, was provided by the NOAA Satellite Active Archive. The the summit pressure time series (Fig. 4c) indicates Natural Sciences and Engineering Research Council of that the 10 May 1996 summit attempt took place Canada provided funding for this research. during a period in which the pressure was falling to values that were significantly lower than those usually observed. This signal, which is indicative of the REFERENCES IHHHHHI large-scale circulation, rather than local wind speed, Adams, J. C., and P. N. Swarztrauber, 1999: SPHERE- which may be influenced by local topographic effects PACK 3.0: A model development facility. Mon. Wea. as well as misleading as the result of the presence of Rev., 127, 1872-1878. jet streaks, should have provided a warning as to the Andersson, E., and Coauthors, 2005: Assimilation and possibility of high-impact weather. The analysis of modeling of the atmospheric hydrological cycle in wind speeds as well as the divergent circulation on the ECMWF forecasting system. Bull. Amer. Meteor. constant pressure or potential temperature surfaces Soc., 86, 387-402. may provide evidence of the presence of jet streaks, Arken, G., 1985: Mathematical Methods for Physicists. shortwave troughs, and their associated ageostrophic 3rd ed. Academic Press, 1029 pp.

478 I BAflS- APRIL 2006 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC Boukreev, A., and G. W. DeWalt, 1999: The Climb: Tragic by the TRMM Precipitation Radar. /. Geophys. Res., Ambitions on Everest. 1st St. Martin's Griffin ed. St. 110, D05106, doi:10.1029/2004JD004815. Martin's Griffin, 372 pp. Horsnell, M., and R. Faux, 1996: Everest blizzard kills Breashears, D., 1999: High Exposure: An Enduring Pas- veteran on fifth climb. The Times of London, May sion for Everest and Unforgiving Places. Simon & 14, Final Edition, 12. Schuster, 319 pp. Hoskins, B. J., 1991: Towards a PV-Theta view of the Burns, J. F., 1996: Everest takes worst toll refusing to general-circulation. Tellus, 43A, 27-35. become stylish. New York Times, May 14, Late Edi- , M. E. Mclntyre, and A. W. Robertson, 1985: On the tion, A1-A3. use and significance of isentropic potential vorticity Challinor, A. J., T. R. Wheeler, J. M. Slingo, P. Q. Crau- maps. Quart. J. Roy. Meteor. Soc., Ill, 877-946. furd, and D. I. F. Grimes, 2005: Simulation of crop Houston, C. S., J. R. Sutton, A. Cymerman, and J. T. yields using ERA-40: Limits to skill and nonstation- Reeves, 1987: Operation Everest-II—Man at extreme arity in weather-yield relationships. /. Appl. Meteor., altitude. /. Appl. Physiol, 63, 877-882. 44, 516-531. Huey, R. B., 2001: Limits to human performance: Elevated Danielsen, E. F., 1984: Meteorological context for the risks on high mountains. /. Exp. Biol., 204, 3115-3119. Global Tropospheric Experiments instruments tests. , andX. Eguskitza, 2000: Supplemental oxygen and Eos, Trans. Amer. Meteor. Soc., 65, 502-518. mountaineer death rates on Everest and K2. /. Amer. Davies, T. D., and E. Schuepbach, 1994: Episodes of high Med. Assoc., 284, 181. ozone concentrations at the Earth's surface resulting , and R. Salisbury, 2003: Success and death on from transport down from the upper troposphere Mount Everest. Amer. Alpine J., 45, 1-10. lower stratosphere—A review and case-studies. , X. Eguskitza, and M. Dillon, 2001: Mountaineer- Atmos. Environ., 28, 53-68. ing in thin air. Hypoxia: From Genes to the Bedside, Devlin, R. B., and Coauthors, 1996: Time-dependent R. C. Roach, P. D. Wagner, and P. H. Hackett, Eds., changes of inflammatory mediators in the lungs Kluwer Academic, 225-237. of humans exposed to 0.4 ppm ozone for 2 hr: A Kalnay, E., and Coauthors, 1996: The NCEP/NCAR comparison of mediators found in bronchoalveolar 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., lavage fluid 1 and 18 hr after exposure. Toxicol. Appl. 77, 437-471. Pharmacol., 138, 176-185. Kayes, D. C., 2004: The 1996 Mount Everest climbing Dickinson, M., 1999: The Other Side of Everest: Climb- disaster: The breakdown of learning in teams. Hum. ing the North Face through the Killer Storm. 1st ed. Rel, 57, 1263-1284. Times Books, 233 pp. Kennett, E. J., and R. Toumi, 2005: Himalayan rainfall Doswell, C. A., and L. F. Bosart, 2001: Extratropical and vorticity generation within the Indian sum- synoptic-scale processes and severe convection. mer monsoon. Geophys. Res. Lett., 32, L04802, Severe Convective Storms, C. A. Doswell, Ed., Amer. doi:10.1029/2004GL021925. Meteor. Soc., 561. Keyser, D., and M. A. Shapiro, 1986: A review of the Environmental Protection Agency, 1996: Air quality structure and dynamics of upper-level frontal zones. criteria for ozone and related photochemical oxi- Mon. Wea. Rev., 114, 452-499. dants. EPA Rep. EPA/600/P-93/004a-c.F, Office of Kidwell, K. B. Ed., 1998: NOAA Polar Orbiter data user's Research and Development, National Center for guide. NOAA/NESDIS, 412 pp. Environmental Assessment. Krakauer, J., 1999: Into Thin Air: A Personal Account of Gruen, A., and S. Murai, 2002: High-resolution 3D the Mount Everest Disaster. Anchor Books, 332 pp. modelling and visualization of Mount Everest. ISPRS Krismnamurti, T. N., 1961: The subtropical jet stream J. Photogramm. Remote Sens., 57, 102-113. of winter. /. Meteor., 18, 172-191. Hemmleb, J., L. A. Johnson, and E. R. Simonson, 2001: Laing, A. G., and J. M. Fritsch, 1993: Mesoscale convec- Ghosts of Everest: The Search for Mallory & Irvine. tive complexes over the Indian monsoon region. /. Mountaineers Books, 203 pp. Climate, 6,911-919. , E. R. Simonson, and D. Hahn, 2002: Detectives on Lang, T. J., and A. P. Barros, 2004: Winter storms in Everest: The 2001 Mallory & Irvine Research Expedi- the central Himalayas. /. Meteor. Soc. Japan, 82, tion. Mountaineers Books, 240 pp. 829-844. Hillary, E., 2003: High Adventure. Oxford University Lau, M. B., 1998: GEOPAK: Monitoring climbers and Press, 245 pp. climate on Mount Everest. Department of Electrical Hirose, M., and K. Nakamura, 2005: Spatial and diurnal Engineering and Computer Science, Massachusetts variation of precipitation systems over Asia observed Institute of Technology, 42 pp.

AMERICAN METEOROLOGICAL SOCIETY APRIL 2006 BAF15- | 479 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC MacGillvray, G., and D. F. Breashears, 1998: Everest. and north-westerly flow. Quart. J. Roy. Meteor. Soc., Mirimax Home Entertainment, 45 min. 125, 2535-2562. Mangione, L., 2003: The 1996 Mount Everest tragedy: Simmons, A. J., and J. K. Gibson, 2000: The ERA-40 Contemplation on group process and group dynam- Project Plan. ERA-40 Project Report Series, No. 1, ics. Int. J. Group Psychother., 53, 353-373. ECMWF, 63 pp. May, W., 2004: Simulation of the variability and ex- Smirnov, V. V., and G. W. K. Moore, 1999: Spatial and tremes of daily rainfall during the Indian summer temporal structrure of atmospheric water vapor monsoon for present and future times in a global transport in the Mackenzie River basin. /. Climate, time-slice experiment. Climate Dyn., 22, 183-204. 12,681-696. Messner, R., 1999: Everest: Expedition to the Ultimate. Spengler, J. D., and D. G. Wilson, 2003: Air quality in Mountaineers, 276 pp. aircraft. Proc. Inst. Mech. Eng. E J. Process Mech. , and T. Carruthers, 2001: The Second Death of Eng., 217, 323-335. George Mallory: The Enigma and Spirit of Mount , S. Ludwig, and R. A. Weker, 2004: Ozone exposures Everest. 1st U.S. ed. St. Martin's Press, 205 pp. during trans-continental and trans-pacific flights. Moore, G. W. K., 2004: Mount Everest snow plume: Indoor Air, 14 (7), 67-73. A case study. Geophys. Res. Lett., 31, L22102 Uccellini, L. W., and D. R. Johnson, 1979: The coupling doi:10.1029/2004GL021046. of upper and lower tropospheric jet streaks and im- , and J. L. Semple, 2004: High Himalayan me- plications for the development of severe convective teorology: Weather at the South Col of Mount storms. Mon. Wea. Rev., 107, 682-703. Everest. Geophys. Res. Lett., 31, L18109, doi:10.1029/ Venables, S., 2003: Everest Summit of Achievement. 2004GL020621. Royal Geographical Society, 252 pp. Morgan, M. C., and J. W. Nielsen-Gammon, 1998: Using Viezee, W., W. B. Johnson, and H. B. Singh, 1983: Strato- tropopause maps to diagnose midlatitude weather spheric ozone in the lower troposphere. Assessment systems. Mon. Wea. Rev., 126, 2555-2579. of downward flux and ground-level impact. Atmos. National Research Council, 2002: The Airliner Cabin Environ., 17, 1979-1993. Environment and the Health of Passengers and Crew. West, J. B., 1984: Human physiology at extreme altitudes National Academy Press, 326 pp. on Mount Everest. Science, 223, 784-788. Parthasarathy, B., A. A. Munot, and D. R. Kothawale, , 1999: Barometric pressures on Mt. Everest: New 1994: All-India monthly and seasonal rainfall se- data and physiological significance. /. Appl. Physiol., ries—1871-1993. Theor. Appl. Climatol., 49, 217-224. 86, 1062-1066. Peixoto, J. P., and A. H. Oort, 1992: Physics of Climate. , 2000: Human limits for hypoxia—The physiologi- American Institute of Physics, 520 pp. cal challenge of climbing Mt. Everest. Ann. New York Ravetta, F., and G. Ancellet, 2000: Identification of Acad. Sci., 899, 15-27. dynamical processes at the tropopause during the , P. H. Hackett, K. H. Maret, J. S. Milledge, R. M. decay of a cutoff low using high-resolution airborne Peters, C. J. Pizzo, and R. M. Winslow, 1983a: Pulmo- lidar ozone measurements. Mon. Wea. Rev., 128, nary gas-exchange on the summit of Mount Everest. 3252-3267. /. Appl. Physiol, 55, 678-687. Roberto, M. A., 2002: Lessons from Everest: The interac- , S. Lahiri, K. H. Maret, R. M. Peters, and C. J. Pizzo, tion of cognitive bias, psychological safety, and sys- 1983b: Barometric pressures at extreme altitudes tem complexity. Calif. Manage. Rev., 45, 136-158. on Mt Everest—Physiological significance. /. Appl. Roelofs, G. J., and Coauthors, 2003: Intercomparison Physiol, 54, 1188-1194. of tropospheric ozone models: Ozone transport in a Wittels, P., M. Neuberger, H. Hauck, and B. Kanduth, complex tropopause folding event. /. Geophys. Res., 1997: Decrease of lung function parameters after 108, 8529, doi:10.1029/2003JD003462. physical exercise in the Austrian Alps at ozone con- Rotunno, R., W. C. Skamarock, and C. Snyder, 1994: An centrations below the statutory "pre-warning phase." analysis of frontogenesis in numerical simulations of Wiener Klin. Wochenschrift., 109, 321-326. baroclinic waves. /. Atmos. Sci., 51, 3373-3398. Yanai, M., and C. Li, 1994: Mechanism of heating and Schultz, D. M., and C. A. Doswell, 1999: Conceptual the boundary layer over the Tibetan Plateau. Mon. models of upper-level frontogenesis in south-westerly Wea. Rev., 122, 305-323.

| BANS' APRIL 2006 Unauthenticated | Downloaded 09/24/21 01:25 PM UTC