Tree wells are close to zero risk on a groomed piste and a real one on a powder day in the trees. The most quoted figure in this field says 72 percent of people who died were skiing with a partner who was not watching. That number is press-relayed and we could not verify who counted it, but the failure it describes is what both deepsnowsafety.org and Austria's KFV build their guidance around. Here is how to turn it into a protocol your kids can actually follow.
Nobody mentions tree wells at the rental shop. They are not on the piste map, not in the lift pass small print, and not in the ski school welcome talk. Then a news story turns up in January, and you spend the rest of the week watching your kids vanish into the trees at the edge of the run.
So, the answer first. For most family ski trips, this hazard is close to zero. It needs deep, loose, recent snow, and it needs trees. On a groomed piste, in a thin snow year, or above the treeline, it is not the thing to worry about, and we will say exactly where the line sits.
But on a powder day at a treed resort, one idea should change how your family skis. The most widely repeated statistic in this field says that 72 percent of people who died in snow immersion suffocation incidents were skiing or riding with a partner and did not have visual contact at the moment it happened. Handle that number carefully, because we are going to be honest with you about it: we found it relayed by Steamboat Pilot & Today, credited to something called the Northwest Avalanche Institute, with no period and no denominator published, and we could not confirm that body as an established research organisation. We are including it because the failure it describes is what both deepsnowsafety.org and Austria's KFV build their guidance around. If it is right, skiing together did not save those people. Watching each other would have.
A tree well is the hollow of loose, unconsolidated snow that forms around the base of a conifer after heavy snowfall, hidden by the low branches. Fall into one, usually headfirst, and the snow behaves like something between quicksand and wet sand: the harder you struggle, the more it packs around you.
Paul Baugher, a ski patrol director who runs deepsnowsafety.org, has logged every publicised in-bounds case in the United States since 2001. His tally is roughly 90 deaths at US ski areas since 2001, an average of four to five a season. Treat that as one practitioner's count of publicised cases rather than an official statistic, and note that we reached it through an Outside magazine writeup of his data, because the site's own numbers page is published as images we could not read. In research experiments recreating these immersions, 90 percent of participants could not get themselves out unassisted, according to deepsnowsafety.org, the resource the National Ski Areas Association points to on this subject. The same source is blunt about the speed: without a partner immediately present, death can come as fast as drowning.
Read those together and the whole article writes itself. You almost certainly cannot self-rescue. The clock is the one that applies to a drowning, not the one that applies to a phone call. Therefore the only protection that works is a person who saw you go in. Not a person nearby. A person who saw.
Snow does not settle evenly around a big spruce or fir. The branches catch and shed it outward, so the trunk sits in a void of deep, airy, uncompacted snow while the ground a metre away is firm enough to hold you. From above, on a fresh morning, the two look identical. That is the entire trick of it.
Austria's KFV (Kuratorium für Verkehrssicherheit), the state-mandated accident prevention body, describes the same formation in German as Baumtrichter or Baumlöcher, deep cavities that form mostly around conifers and dwarf pines after heavy snowfall and are frequently concealed by low branches. KFV also makes the point that matters most: where new loose snow gives sink depths of a metre or more, a fall can put you in deep enough that you cannot free yourself, and the instinct to fight your way out makes it worse.
deepsnowsafety.org's own case tally attributes roughly 70 percent of these accidents to tree wells specifically, with the rest in open deep snow away from trees. That proportion comes from the organisation's own logbook rather than an independent registry, so treat it as directional. The practical version: trees concentrate the hazard, but deep loose snow alone is enough.
Real talk on the numbers. The counts in this article are compiled from publicised in-bounds cases by a safety practitioner, not from a government death registry, and the underlying datasets are not published in a form we could inspect line by line. They are the best figures that exist, and they are almost certainly a floor rather than a ceiling. We would rather tell you that than pretend the sourcing is tidier than it is.
We have already told you what is shaky about that number. What is not shaky is the failure mode it points at, because it is what both deepsnowsafety.org and Austria's KFV build their deep snow guidance around: KFV's own list says in plain terms to keep mutual visual contact in a group. That is prescriptive advice rather than a second count of victims, so we are not going to dress it up as independent confirmation of the 72 percent itself.
Here is the thing. Almost every parent already believes they have solved this problem, because they solved it years ago at the swimming pool and the car park and the ski hill: we stay together. Nobody skis alone. Everybody knows where everybody is.
That belief is where families get caught. The people in that statistic were not solo. They had partners. They had a plan that sounded exactly like yours. What they did not have, in the moments that decided it, was a pair of eyes on them.
In trees this happens without anyone being careless. Your 13-year-old drops into a glade ten metres ahead. You follow the line she took, you lose her behind a stand of firs, you assume she is below you, and you keep going. That is the entire failure mode. There is no shout, no crash, no signal. Deep snow is quiet; a person who goes in headfirst makes almost no noise at all, and their skis or board do not stay on the surface waving like a flag. That is reasoning from how loose deep snow behaves, not a finding we could source, but it is why the leapfrog rule below is built on eyes rather than ears.
The good news is that this is one of the few mountain risks with a clean, cheap, teachable fix. It does not require gear, fitness, or expertise. It requires converting we ski together into we can see each other right now, and drilling that until a child can say it back to you.
| Where you are | Is this a real risk? | What to do about it |
|---|---|---|
| Groomed, marked piste at any resort | No. Effectively zero. | Nothing. Ski. This hazard needs deep unconsolidated snow, and a groomed piste by definition does not have any. |
| Above the treeline, on open high Alpine terrain | No trees, so no tree wells. | Other hazards live up there (whiteout, wind, avalanche terrain off-piste), but not this one. If your day is spent above the trees, you can put this particular worry down. |
| Thin snow year, firm or spring snow, an icy Alpine February | Very low. | Tree wells need loose new snow deep enough to sink into. Settled, wind-packed or refrozen snow does not swallow a skier. |
| Marked run with trees along the edge, after a heavy snowfall | Low, but not zero if a child cuts a corner into the trees. | Ski the middle third of the run. Make "stay off the tree side today" a rule for the day rather than a suggestion. |
| Glades and tree runs during a storm, or while the new snow is still loose and deep | Yes. This is the one. | Full pair protocol, whistles, continuous visual contact, or do not go into the trees at all that day. |
| Tree runs at a high snowfall resort on a big day | Highest family exposure, and it is usually the teenager rather than the toddler. | Have this conversation before the trip, not at the top of the run. If Japan is the trip you are planning, start with our Japan family guide. |
| Beginner and intermediate ski school groups on marked pistes | No. | Those lessons run on marked, patrolled terrain, which is not where these incidents happen. Advanced lessons, off-piste clinics and guided tree sessions are a different product, so ask what the terrain will be when you book. |
The honest picture is not the one the phrase "family ski danger" conjures. KFV logged six deep snow suffocation deaths in Austria between 2018 and 2021, and the victims it describes are adults aged roughly 24 to 39, including a 39-year-old skier and two snowboarders in their twenties. No children are named in that list, and we are not going to imply otherwise.
We looked for a dataset breaking snow immersion victims out by age and could not find one. So we will tell you what we think rather than dress an inference up as a statistic: the family member most exposed to this is not your five-year-old, it is your competent 12 to 16-year-old. Small children ski where you put them, which is the piste. Teenagers ski where their friends go, which on a powder day is into the trees, half a run ahead of you, on the day the snow is deepest and the visibility is worst.
That reframes what the conversation is for. You are not warning a small child away from a monster in the woods. You are giving a confident adolescent a rule that survives peer pressure, because "my mum said no trees" does not, and "we go in pairs and we leapfrog" does. It sounds like what good skiers do, because it is what good skiers do.
What does not help, so do not buy it. An avalanche transceiver does not solve this problem. The failure here is not "we could not find them", it is "nobody knew to look". A beacon, a RECCO reflector, or an airbag pack all assume someone is already searching within minutes. If your protocol produces that someone, you did not need the hardware. If it does not, the hardware arrives too late.
Parents reach for the phone answer immediately, and it is worth explaining precisely why it fails, because the reasoning is what makes the alternative stick.
Start with the speed. The deep snow safety guidance we can actually source does not hand you a countdown in minutes, and we are not going to invent one. What it says is that without a partner immediately present, death can occur as fast as drowning. Now walk through the actual sequence a phone sets in motion: someone notices an absence, backtracks, fails to find them, calls the resort, explains a location in trees with no piste marker, and patrol begins a search of a wooded slope in falling snow. Hold that against the pace of a drowning and you have your answer.
Then there is the buried person's end of the call. In a headfirst immersion your arms are typically pinned below or beside you and you are inverted in loose snow. Getting a gloved hand into a zipped inside pocket, extracting a phone, and operating a touchscreen upside down is not a realistic sequence. We would not build a plan around a watch's fall detection either: a soft, silent, slow sink into powder is not the sort of event those features are looking for, and we have seen no evidence that anyone has tested them against it.
Carry the phone. Call patrol the second you have someone else on scene, and know the local emergency number for the country you are in. But the phone is what you use after the rescue has already started, and the rescue starts with the person who was watching.
This is information for your family to know in advance, not something anyone works out mid-fall. Teach it on the chairlift.
Grab something on the way down. If you feel yourself going in, branches, the trunk, anything. Stopping the fall at chest height is a completely different situation from ending up inverted.
Do not thrash. This is the counterintuitive one and it is why it must be taught in advance. KFV's guidance is explicit that loose deep snow behaves in a way that makes violent self-rescue attempts worse: struggling packs snow around you and consumes the space and air you have. Stay calm is not soothing advice here, it is mechanical advice.
Protect the airway first, always. Before trying to move your body, use whatever hand movement you have to make and keep a space in front of your mouth and nose. Everything else is negotiable; that is not.
Make noise the way that works. Whistle. KFV's own list for deep snow includes exactly this, keep calm and use a whistle to attract attention. Shouting from under snow carries poorly and burns air.
Then move slowly, toward the light and away from the trunk. Deliberate, small, rocking movements to enlarge the space, not a fight. If your feet are above you and your skis are still on, they are anchors; getting them off is worth the effort if you can reach the bindings.
KFV also recommends a full-face helmet visor or a balaclava for deep snow riding, which helps keep loose snow out of the airway. In practice a balaclava or a well-fitted neck gaiter is the version a family will actually wear all day, and it is a reasonable habit on a storm day.
Mistake one: skiing off to get help. Every instinct says find a grown-up, find patrol, find the lift attendant. In this specific scenario that instinct kills, and it is worth saying that plainly to your children in those words. With nine out of ten people in those research recreations unable to free themselves, and a timescale the sourcing compares to drowning, leaving means they spend the whole of it alone. Your partner is the rescue, and that conclusion comes straight out of those two numbers rather than out of any slogan. Stay, dig, and shout for help from where you are. If there are three or more of you, one person can go for patrol while the rest keep digging.
Mistake two: pulling on the legs. If you can see boots or skis sticking up, the reflex is to haul. We could not find the rescue mechanics in the sources we verified for this article, so take the next two sentences as reasoning from how loose snow behaves rather than as an instruction from an authority. The head is the deepest point, so hauling on the lower body works against the only air space the person has, and it drags them through exactly the loose snow that packs. Dig to the face instead. Go in from the side, at an angle, and get snow away from the head and chest first. Talk the whole time so they know someone is there, which does more for a buried person's panic than anything else available. Ask patrol at your resort what they teach for this; a local, current answer beats ours.
Mark the spot the moment you see it, with a pole or a jacket, because a slope of trees looks identical from three metres away and people lose the location while going for a shovel.
This is safety information, not medical advice. Anyone who has been immersed, even briefly, even if they get up and laugh about it, should be seen by ski patrol and then by a doctor. Snow inhalation and cold exposure both have delayed presentations, and a child in particular may not report feeling wrong.
In almost all realistic family situations, no. Tree wells need deep, loose, recent snow and conifers, which is off-piste or gladed terrain, not where a six-year-old skis. Beginner and intermediate ski school groups operate on marked, patrolled pistes. The family member with meaningful exposure is a confident older child or teenager who follows friends into the trees on a powder day.
They happen in Europe. Austria's KFV (Kuratorium für Verkehrssicherheit), the country's state-mandated accident prevention body, publishes guidance on deep snow suffocation and logged six deaths in Austria between 2018 and 2021 (victims aged roughly 24 to 39). The hazard is simply discussed more in North America and Japan than it is in Europe. Why that is, we can only guess at, and we are not going to pass a guess off as a fact. If you ski Austrian or Swiss tree runs after a big dump, the same rules apply.
No. A groomed piste has compacted, machine-worked snow that will not swallow a skier, and there are no conifer trunks in the middle of it. The risk lives in the untouched snow at the sides and in the glades. This is why we say plainly that most family ski weeks have effectively zero exposure to this hazard: telling you when a fear does not apply is the point.
We are not going to give you a number of minutes, because the sources we can verify do not give one and inventing a countdown would be worse than admitting the gap. What deepsnowsafety.org says is that without a partner immediately present, death can occur as fast as drowning, and that in research experiments recreating these immersions 90 percent of participants could not extricate themselves without help. That combination is why the response has to come from someone already standing there, not from a phone call.
Not for this. A transceiver, a RECCO reflector and an airbag all assume that someone already knows a person is missing and is actively searching within minutes. The failure here is that nobody knew to look, which is exactly what the most quoted figure in this field describes, with the caveat that we could not verify who counted it. Fix the watching, and the hardware becomes unnecessary. Skip the watching, and the hardware arrives too late. A whistle is the one purchase worth making.
Give them a sentence, not a briefing: "If I cannot see you, I stop and I shout." Then name their partner out loud before every treed run, and make stopping always the right call, even when it turns out you were just behind a tree. Children retain lines and roles. They do not retain risk statistics, and you should not be reciting fatality numbers to a seven-year-old.
Partly, and we would rather say so. The count of roughly 90 in-bounds US deaths since 2001, averaging four to five a season, is Paul Baugher's own logbook of publicised cases at deepsnowsafety.org, and we reached it through an Outside magazine writeup rather than an official registry. The 72 percent no-visual-contact figure is weaker still: we found it in Steamboat Pilot & Today, credited to the Northwest Avalanche Institute, with no period or denominator published, and we could not confirm that body as an established research organisation. Read both as directional and as a floor rather than a ceiling. The mechanism your protocol is built on is not in dispute, and is described the same way by deepsnowsafety.org in the US and KFV in Austria.
Recevez notre check-list gratuite pour skier en famille et un e-mail utile par mois.
Utilisez nos outils pour trouver la station idéale pour votre famille.
Unguided unmarked terrain with kids is not a family product. What Italy, Switzerland, France, and Austria actually say, and which ESF courses start at 13, 14, or 18.
How-To GuidesEverything you need to pack for a stress-free family ski vacation, organized by timeline and age group.
How-To GuidesEvery option for getting from Milan to Cortina with children. Trains, buses, rental cars, and private transfers compared.
Note de transparence : Ce contenu a été créé avec l'aide de l'IA et relu par Tom Meredith, notre rédacteur. Les prix, dates et disponibilités peuvent changer. Nous recommandons de confirmer les informations directement auprès de la station avant de réserver.