On a backcountry ski tour outside Koktokay, our guide stopped on an open meadow and said, “Today you build a quinzee.” I had read about snow shelters in expedition books but never made one. Two hours later I was curled inside a dome of packed snow at minus 18, warmer than the tent I usually carry. Here is the method, and why it now matters to every trip I take into the quiet country.
Why a quinzee, not a snow cave
A wilderness first-aid course had taught me that snow caves need a slope with deep, stable snow — and that is rare on the open Altay meadows where most of our tours run. A quinzee is different: you pile loose snow into a mound, let it sinter for an hour, then hollow it out from the inside. No slope required. For a avalanche-aware trip, that flexibility is the whole point, because you can build one wherever the snow is deep enough to pile, even on flat ground that would never give you a cave.
Tools you actually need
A wide snow shovel, a small saw or pruning blade for the vent, and a long stick or two to gauge wall thickness. That is it. I now carry a folding shovel that clips to my pack, because the one time I borrowed a guide’s and then skied away without my own, I spent an anxious night wondering if I could build a second one with just my skis. You do not need a tent, and you do not need a saw unless you want a clean vent hole.
Step one: pile high and mix your layers
We shoveled a cone about 2.2 metres tall and 3 metres wide at the base. The trick the guide drilled into me was to mix layers as I piled — fresh powder on top of packed, then stomp it, then more powder. You do not want a solid ice block; you want settled, bonded snow that insulates. We stuck a few twigs halfway into the surface as depth gauges so neither of us would dig through the thin roof later in the dark, when judgement is worst.
Step two: wait, and do not rush it
Snow needs time to bond into a shell. We skied a short lap on a nearby ridge for about 60 to 90 minutes while the mound settled. Rushing this step risks a collapse onto your head, and I have seen a half-built quinzee slump when someone tried to hollow it too early. On a guided off-piste day this wait is easy to fit in; solo, I treat it as non-negotiable and use the time to melt snow for water, which I will need later anyway.
Step three: hollow carefully from a low entrance
Crawl in from a low entrance and scoop upward and outward, staying at least 30 centimetres from the surface. We made a small vent with a ski pole through the top so air could move. Inside, the temperature held near zero while outside it was minus 18. The silence was the part I did not expect — a snow dome muffles wind completely, and I could hear my own heartbeat, which was oddly calming after a day of exposed ridge lines and gusting summits.
A real emergency that justified the skill
A season later, a partner twisted a knee on a remote loop and we could not move before dark. The quinzee I built that evening kept us both warm until help reached us at noon. I am not romantic about it — it was cold and uncomfortable — but it turned a scary night into a survivable one, and that is the entire reason to learn the skill before you need it rather than during.
When I actually use it, and when I do not
A quinzee is not for every trip. On a resort day it is pointless, and on a tight itinerary it eats too much time. But for an extended winter expedition or an unplanned bivouac, it is lighter than a tent and warmer than sitting in the open. I now practise once a season on a safe meadow so the muscle memory is there, because nobody learns a shelter skill well under pressure, with cold fingers and a fading light.
Honest limits and safety non-negotiables
It is slow. Building a solid quinzee takes two people and a couple of hours of focused work. If you are mobile and the weather is clear, keep moving toward safety rather than building. And never sleep in one without a vent — carbon dioxide builds fast in a sealed snow dome, and that is a silent failure mode. I always leave the entrance partly open and dig a breathing channel before I rest, and I check my partner is awake before I let myself doze.
I keep a note of shelter techniques in my skiing notes and review them before any backcountry trip. It is a small skill that buys a lot of peace of mind, and the first time you sleep warm in a hole you dug yourself, the confidence sticks for good.
Choosing a build site
Pick a flat spot with deep, unpacked snow and no overhead hazard — no trees that could drop a load of snow, no slope above that could avalanche. We built ours on a meadow a safe distance from any path. Avoid shaded hollows where the snow is sugar-loose and will not bond. The guide taught me to test by squeezing a handful: if it clumps and holds, it will sinter; if it falls apart, find elsewhere. Thirty minutes of site choice saves two hours of a collapsing dome.
Teaching a child the skill
On a later trip I showed my nephew, then eleven, how to build one, and it became the best outdoor lesson he had ever had. Kids love the pile-and-hollow game, and the payoff — a secret snow room — is immediate. We kept it to a daytime practice shelter, never an overnight, and I supervised the hollowing so he did not dig a thin roof. The confidence he gained around snow was worth more than any classroom session about winter, and he now checks ice and snow on his own.
Quinzee build checklist
Deep unpacked snow; a wide shovel; two depth-gauge sticks; an hour of wait time; a partner; and a vent pole. Tick every box before you start, because a quinzee built without the wait or without a vent is a hazard, not a shelter. I keep a printed card of these steps in my pack now, because cold and fatigue make even simple sequences slippery, and the card removes the temptation to skip the boring parts.
When NOT to build one
If you are mobile and the weather is clear, keep moving — a quinzee is for when you must stop, not a novelty. If the snow is shallow or sugary and will not bond, do not waste the hour; build a wind wall and move on. And never build one alone on a big trip without telling someone your location and expected time. The skill is a tool, and like every tool it has a wrong time to use it, which is most of the time.