The Physician and Sportsmedicine
Menubar Home Journal Personal Health Resource Center CME Advertiser Services About Us

Eight Nonavalanche Snow-Immersion Deaths

A 6-Year Series From British Columbia Ski Areas

Robert Cadman, PhD


In Brief: The non-avalanche-related snow immersion death (NARSID) appears to be an emerging fatality pattern in Alpine skiing and snowboarding in western North America. Of 32 skiing fatalities that occurred in British Columbia from 1993 to 192021, 8 were NARSIDs. Six deaths occurred when victims suffocated after falling into tree wells around conifers; two victims asphyxiated after falling into a deep snowbank. Seven were Alpine skiers and one was a snowboarder; all had advanced skiing or snowboarding skills. All recorded NARSIDs occurred during or immediately after heavy snowstorms. The patterns found here suggest prevention strategies and show the importance of using avalanche rescue dogs as early as possible in the search for missing skiers and snowboarders.

Mortality in Alpine skiing is well documented, with a rate of 0.2 to 0.7 deaths per million skier visits (1-9). Fatalities usually result from collision with a tree or rock inside a ski area or asphyxiation caused by an avalanche outside ski area boundaries (1-3,10). Among snowboarders, the fatality rate is not well documented, but anecdotal evidence suggests that the rates are similar to those of Alpine skiers. Shealy (8) reported that the fatality rate for snowboarders in the United States is 0.46 per million ski-area visits.

Snow-immersion deaths in Alpine skiing and snowboarding are usually associated with avalanche burials in the backcountry or in unpatrolled regions outside ski areas but accessed from ski lifts (2,10,11). Non-avalanche-related snow-immersion deaths (NARSIDs), often called tree-well deaths, have not been well documented. Such deaths occur when an Alpine skier or snowboarder falls, usually headfirst, into a deep space around a tree or in a deep-powder drift and suffocates (figure 1: not shown).

In 1994, Kizer et al (12) reported three NARSIDs among snowboarders that occurred during a severe storm cycle in the California mountains in 1992. The three young men were found in wooded areas off the designated ski trails, head-down in loosely packed snow at the bases of trees. In each case, the bottom of the snowboard was visible at the snow surface and no evidence of collision with the tree was found. Another NARSID was reported by Shephard in 1996 (13). He described an 11-year-old male Alpine skier who was found buried headfirst in a meter of snow with no signs of a struggle or collision.

These reports, along with my experience as a ski patroller who has located two NARSID victims, prompted this investigation.


A retrospective case-series analysis of coroner's reports from January 1, 1993, to December 31, 192021, was obtained. The computer database of the British Columbia chief coroner includes reporting codes for deaths related to Alpine skiing and snowboarding. After cases were identified by a computer search for skiing and snowboarding deaths, the hard-copy files were reviewed. The analysis focused on the files of victims who were asphyxiated from non-avalanche-related snow immersion and had been skiing or snowboarding within a ski-area boundary or in areas just outside the boundary but accessed from the ski area.

Files included the autopsy report, the police report, the ski patrol report, and, when applicable, the coroner's report and/or inquest. To ensure that no files were missed, the coroner's data were cross-referenced with ski industry data. The data were coded and analyzed using Statistical Package for the Social Sciences.


In the 1993 to 192021 study period there were 32 trauma fatalities, excluding avalanche and lift-related fatalities, at British Columbia ski areas. Ski-injury researchers generally exclude avalanche- and lift-related deaths. Avalanche deaths are excluded because they generally occur outside the ski area on slopes that have not been controlled for avalanche hazards. Lift-related deaths are usually caused by mechanical failure and are beyond the control of the skier. During the study period there were no avalanche deaths within British Columbia ski areas and three lift-related deaths.

Ten of the of the 32 trauma fatalities were from asphyxiation. Eight cases involved snow immersion and are included in this analysis. Excluded cases were a victim who drowned in water and another who aspirated vomitus while unconscious.

The Canada West Ski Areas Association reported 26,000,000 Alpine skier and snowboarder visits during the study period (personal communication, J. Spencer, April 192021). The incidence rate of NARSIDs during the study period was 0.31 per million skier visits.

As indicated in table 1, seven victims were male Alpine skiers. Five of the eight victims were skiing within the ski area boundary but off designated trails. Out-of-bounds Alpine skiers and snowboarders accounted for 3 of the fatalities. The mean age of the deceased was 28 years (range, 9 to 49 years).

Table 1. Non-Avalanche-Related Snow-Immersion Deaths in British Columbia, 1993-192021

Case Date Age Gender Location Activity Mechanism
1 Jan 93 33 M IB trees Skiing Tree well
2 Dec 94 22 M IB trees Skiing Tree well
3 Feb 94 30 M OB trees Skiing Deep snow
4 Mar 95 9 F IB trees Skiing Tree well
5 Dec 96 42 M IB open terrain Skiing Deep snow
6 Dec 96 19 M IB trees Skiing Tree well
7 Dec 96 49 M OB trees Skiing Tree well
8 Feb 2021 23 M OB trees Snowboarding Tree well

IB = in bounds (within ski area boundary); OB = out of bounds

Tree wells were the most common feature of the fatalities; six occurred in soft snow around the base of coniferous trees (figure 2), and two cases involved deep snow without a tree well. The victim in case 3 was skiing and fell headfirst into a deep snowdrift and could not free himself. The subject in case 5 was an advanced skier who became wedged under a fallen tree beside an expert-only run during a severe snowstorm. He dislocated his shoulder in the fall, was unable to free himself, and was asphyxiated in the snow. None of the other victims had evidence of trauma in their falls.

[Figure 2]

Three victims (cases 1,5,6) were skiing alone at the time of death; four (cases 2,3,7,8) were skiing in groups of two to four people, and one (case 4), a 9-year-old girl, was with her ski school class of more than 10 children and an instructor. It is thought that she may have jumped alongside an intermediate ski run and landed off the designated run in a tree well. No one in her class saw what happened. The ski area immediately initiated a search, and several hours later an avalanche rescue dog was brought into the search. The dog team followed the path between where the girl had last been seen and where the instructor noticed her absence (approximately 150 m [492 ft]). Within minutes the dog found the girl submerged in a tree well.

An avalanche rescue dog was also instrumental in finding the victim in case 8. The snowboarder had slid silently into a deep tree well while his riding partner was slightly ahead and on a different line through the trees. This victim's head was almost 3.0 m (9.8 ft) below the surface, and his snowboard was 0.30 m (.2021 ft) below the surface, making him invisible to searchers. Despite the struggle to free himself and the ensuing search by his riding partner and the ski patrol, an avalanche rescue dog was needed to locate the victim the following day.

In all cases, resuscitation was delayed 15 minutes to 20 hours because the victims were alone or their friends did not witness the fatal fall. In several cases, the victims' friends retraced their paths and found their missing partners, but in other cases group members assumed the missing person had skied off ahead or behind the group and did not report him or her missing until later that day.

All of the deceased were advanced skiers (including the 9-year-old girl). The deaths occurred at large, medium, and small ski areas.

The only consistent contributing factor that surfaced in the analysis was that NARSIDs occurred during or immediately after large snowstorms. Unfortunately, details about the snowpack were absent from the majority of the case files. In the two cases that I personally investigated, the depth of soft snow around the victims was between 1.65 m (5.4 ft) and 2.50 m (8.2 ft). As defined by the Canadian Avalanche Association Manual (14), the snow density was "fist resistance" for the entire depth, but because of the victim's struggle and resulting heat loss, the snow immediately around the victim was "knife resistance."

Toxicology reports were evaluated for each victim because drugs and alcohol sometimes contribute to fatalities in Alpine skiing and snowboarding (15-17). Only one victim (case 7) tested positive for an illicit drug. He had traces of delta-9-tetrahydrocannabinol (THC) in his blood, and the pathologist estimated that marijuana had been consumed 12 to 24 hours before death.


During the past six ski seasons, 8 of 32 fatalities on British Columbia ski slopes were NARSIDs. This finding has not been previously reported and may indicate an emerging pattern.

Kizer et al (12) and Shealy (8) reported that NARSIDs are a particular hazard for snowboarders. In a recent US study that spanned about the same time frame as this investigation, Shealy (8) reported more snowboarders than Alpine skiers among tree-well fatalities. In contrast, seven of the eight NARSID victims in the present study were Alpine skiers.

Readers should be cautious when interpreting this difference. First, these two studies contain data about infrequent events, which can lead to misinterpretation. Without further research examining the exposure of at-risk populations, comparisons should be made cautiously. Second, NARSIDs appear to be a phenomenon of western North America; few, if any, cases have occurred at ski areas in eastern North America. Current national surveillance systems are unlikely to identify this fatality pattern because the cases are few and, when included as part of a data set, they are typically recorded as "other causes."

However, in British Columbia, with its deep snows and heavily timbered slopes, NARSIDs represent a major cluster among fatally injured Alpine skiers and snowboarders.

Prevention Is Imperative

Few prevention efforts have been designed because NARSIDs are rare and an emerging pattern has not been previously described. Prevention efforts are crucial because, in my view, self-rescue is nearly impossible once a skier or snowboarder becomes inverted in a deep-powder tree well or snowbank.

In a carefully controlled experiment during the winter of 192021-1999, volunteer Alpine skiers and snowboarders were placed inverted into a packed snow hole that simulated a tree well. None of the 10 volunteers (6 skiers, 4 snowboarders) could rescue themselves. Alpine skiers who were able to remove their skis by using their feet to undo their bindings only fell deeper into the hole. The snowboarders could not unfasten their boards. We found that most ski clothing acts like a funnel when inverted. All of the volunteers complained of snow falling onto their bare torsos, suggesting that hypothermia could be a major complicating factor.

Ski patrollers must recognize the potential for NARSIDs when searching for a missing skier who has a history of skiing or riding in wooded areas. Trained avalanche rescue dogs, such as those with the Canadian Avalanche Rescue Dog Association, can locate not only avalanche victims but also those who inadvertently become immersed in deep snow or tree wells. Ski patrols should consider revising search policies to allow for the earliest use of rescue dogs, particularly when the missing person has a history of skiing deep powder in unpatrolled treed areas.

The ski industry and medical community should include NARSID prevention information in education initiatives for the general skiing population. (See the Patient Adviser, "How to Stay Alive In Deep Powder Snow," page 44.)

Levine and Gorman (18) describe why it can be difficult to educate the general skiing population about the risks associated with Alpine skiing and snowboarding. The risk of death in Alpine skiing is considerably less than in many other forms of transportation, and because fatalities are infrequent, skiers are often unaware of them unless the deaths involve high-profile individuals (1). Levine and Gorman (18) postulate that skiers who hear about ski fatalities might decrease risk-taking behaviors; however, they found that only 50% of their study respondents had heard about skiing fatalities. This lack of awareness may cause skiers to underestimate the risk of skiing.

Risk awareness is important regardless of skiers' and snowboarders' skill level. Ekeland et al (19) note that powder skiers are likely to be more skilled than those who stay on packed snow. But the present study suggests that expertise in skiing or snowboarding is unlikely to protect athletes from NARSID unless they are aware of the specific risks, since all of the victims had advanced skills.

Further magnifying the importance of education, the level of exposure to tree-well hazards appears to be high, at least among young skiers. This author surveyed 863 young skiers (under age 18) and found that 68% skied at least one run per day in timbered areas (20).

The frequency of exposure to tree-well hazards may also be amplified by the increasingly common practice of "glading" new ski runs to reduce their environmental impact. "Gladed" ski runs are selectively logged (rather than clear-cut) with trees spaced 2 to 3 m (6.6 to 9.8 ft) apart. Ski areas should follow two important precautions: they should remove the lower branches of coniferous trees to prevent tree-well formation, and they should allow enough space between trees to prevent skier collisions. Also, skiers must understand and distinguish risks of skiing gladed runs and skiing among trees off designated ski runs, where the risks are higher because of low-hanging branches and undisturbed snow.

Alpine skiing and snowboarding are great sports enjoyed by many, but efforts to reduce morbidity and mortality should be increased, based on well-known risks such as collisions and avalanches, as well as the lesser-known risks such as non-avalanche-related deep snow immersion.


  1. Shealy JE, Thomas T: Death in downhill skiing from 1976 through 1992—retrospective view, in Mote CD Jr, Johnson RK, Hauser W, et al: Skiing Trauma and Safety: Tenth International Symposium, ASTM STP 1266. Philadelphia, American Society for Testing and Materials, 1996, pp 66-72
  2. Berghold F: Fatal skiing accidents in Austria—epidemiology and analysis, in Johnson RK, Mote CD Jr, Binet MH (eds): Skiing Trauma and Safety: Seventh International Symposium, ASTM STP 1022. Philadelphia, American Society for Testing and Materials, 120219, pp 63-68
  3. Tough SC, Butt JC: A review of fatal injuries associated with downhill skiing. Am J Forensic Med Pathol 1993;14(1):12-16
  4. Ambach E, Tributsch W, Henn R: Epidemiology of fatalities in Alpine skiing—120217-1990 [in German]. Beitrage zur Gerichtlichen Medizin 1992;50:333-336
  5. Morrow PL: Downhill skiing fatalities: the Vermont experience. J Trauma 120218;28(1):95-100 [published erratum in J Trauma 120218;28(4):561]
  6. Morrow P: Ski fatalities: the Vermont experience. Presented at 13th International Congress on Ski Trauma and Skiing Safety, May 2-8, 1999, Breuil Cervenia, Italy
  7. Shealy J: Death in downhill skiing, in Johnson RJ, Mote CD Jr (eds): Skiing Trauma and Safety: Fifth International Symposium, ASTM STP 860. Philadelphia, American Society for Testing and Materials, 120215, pp 249-357
  8. Shealy JE: Modalities of death, snowboarding and skiing differences; 1991/92 through 1997/2021. Presented at 13th International Congress on Ski Trauma and Skiing Safety, May 2-8, 1999, Breuil Cervenia, Italy
  9. Burtscher M, Nachbauer W, Schröcksnadel P: Risk of traumatic death during downhill skiing compared with that during mountaineering, in Johnson RJ, Mote CD Jr, Ekeland A (eds): Skiing Trauma and Safety: Eleventh Volume, ASTM STP 1289. Philadelphia, American Society for Testing and Materials, 1997, pp 23-29
  10. Tough SC, Butt JC: A review of 19 fatal injuries associated with backcountry skiing. Am J Forensic Med Pathol 1993;14(1):17-21
  11. Furrer M, Erhart S, Frutiger A, et al: Severe skiing injuries: a retrospective analysis of 361 patients including mechanism of trauma, severity of injury, and mortality. J Trauma 1995;39(4):737-741
  12. Kizer KW, MacQuarrie MB, Kuhn BJ, et al: Deep snow immersion deaths: a snowboarding danger. Phys Sportsmed 1994;22(12):49-61
  13. Shephard RJ: Asphyxial death of a young skier. J Sports Med Phys Fitness 1996;36(3):223-227
  14. Schaerer P (ed): Observation Guidelines and Recording Standards for Weather, Snowpack and Avalanche Observations (technical document). Canadian Avalanche Association, Revelstoke, BC, 1995
  15. Meyers A, Perrine M, Caetano R: Alcohol use and downhill ski injuries: a pilot study. Presented at 11th International Congress on Ski Trauma and Skiing Safety, April 23-29, 1995, Oslo, Norway
  16. Cherpitel CJ, Meyers AR, Perrine MW: Alcohol consumption, sensation seeking and ski injury: a case-control study. J Stud Alcohol 192021;59(2):216-221
  17. Salminen S, Pohjola J, Saarelainen P, et al: Alcohol as a risk factor for downhill skiing trauma. J Trauma 1996;40(2):284-287
  18. Levine R, Gorman BS: Skiers' perceptions of danger as a function of awareness of fatalities. J Sports Behav 1994;17(1):17-23
  19. Ekeland A, Lystad H, Holtmoen A, et al: Injuries in powder skiing. Scand J Med Sci Sports 1996;6(6):337-340
  20. Cadman RE: Ski Injury Prevention: An Epidemiological Investigation of the Social, Behavioural and Environmental Determinants of Injury, thesis. Vancouver, BC, University of British Columbia, 1996

The author thanks Jimmie Spencer of Canada West Ski Areas Association and Peggy Justiceson of the Office of the Chief Coroner of the Province of British Columbia for their support of this research project.

Dr Cadman is lecturer in the School of Nursing and Public Health at Edith Cowan University in Perth, Western Australia. He worked for 17 years as a ski patroller in British Columbia. Address correspondence to Robert Cadman, PhD, School of Nursing and Public Health, Edith Cowan University, 100 Joondalup Dr, Joondalup, WA, Australia, 6027; e-mail to: [email protected].