ROUTE FINDING AND PROBING ON A GLACIER

The Technique

1. Route planning starts long before the trip, and it starts with the dry glacier. Get late season satellite or aerial imagery of the glacier when the ice was bare, and use it to build a picture of where the crevasse fields sit. Combine that with a topographic map for slope, aspect and the shape of the valley, and with any published route description. Then get the thing that matters most, which is current local knowledge: a guiding operation that works that glacier, a hut warden, a ranger station, a heli operator, someone who was on it this month. Imagery tells you where the crevasses have been. Only a person who has been there recently tells you what the bridges are doing now. If you cannot get that, you do not have a plan, you have a guess.
2. Where an established route exists, use it. Glaciers that see regular traffic often have a known line, sometimes marked with wands, sometimes carried as a GPS track by a guide service. That line exists because someone worked out where the safe ground is, usually over years. Following it is not weakness, it is the entire method. Understand its limits though. A wand line is only as current as the last person who checked it. A GPS track from last week was safe under last week's conditions and last week's temperatures. A track from last season is a hint about the zones, not a route, because the ice has moved since. And a track on a screen has a position error of its own, which on a glacier can be the difference between the middle of a bridge and the middle of a hole. Navigate by ground and by track together, never by track alone.
3. The safest line is very rarely the most direct one. Crevasse zones are localised, so a route that adds distance to stay in the middle of a uniform flow, away from the margins, off the convex rolls and clear of the ground around a rock outcrop, is doing real work. Where you must cross a crevasse field, cross it at right angles to the crevasses rather than travelling along them, because a line parallel to a hidden slot can sit on top of it for a long way. Contour around a roll rather than driving over the crest of it. Accept a longer, flatter, duller line as the price of the terrain. If the party's mood is pushing you toward the direct line, that is a signal to stop and reassess, not to hurry.
4. Learn what a bridged crevasse looks like from the surface, while accepting that many show nothing. The classic sign is a shallow linear sag, a slight dishing in an otherwise even surface, often only visible when the light rakes across it. Look for faint straight or gently curving lines, a change in snow texture or colour along a line, small collapse holes strung out in a row, or a crack at what would be a lip. Low sun early and late reveals this ground. Flat light and cloud erase it completely, which is why flat light on a glacier is a stop condition rather than an inconvenience: in flat light you lose the only surface information you had, and you are navigating on planning alone.
5. Probing is how you test a specific point, and it has a specific technique. Use a proper avalanche probe, or a pole with the basket removed, and push it in vertically. Vertical matters, because an angled probe can slide along the far wall or the underside of a bridge and give you a completely convincing solid reading over a hole. Probe with steady pressure and read the resistance: what you are feeling for is a sudden loss of resistance, a void, or a distinct change in the layering compared to the ground you know is good. Probe at intervals close enough that you cannot step over a gap between probes, and probe a line across the suspect ground rather than a single hopeful poke. Anyone probing on a glacier should be protected while doing it, which in a mountaineering party means belayed from good ground.
6. Now the honest limits of probing, because this is where people fool themselves. A probe tells you about one point, at one moment, to the depth of the probe. It does not tell you the strength of the bridge, only that something is there. It can find a false floor: a wedged block, a lower bridge, or old debris sitting inside the crevasse, all of which feel like ground. It cannot be done at riding speed, and probing an entire route is not practical over any real distance. Most importantly, a probe measures resistance to a thin rod pushed by a person. It says nothing about whether the same bridge will carry a machine. Probing is a tool for checking a specific suspect spot, a stopping place or a landing zone. It is not a way to make a glacier safe.
7. Time your travel around cold. Bridges are strongest after a long, cold, clear night with a hard refreeze, and they weaken through the day as air temperature, sun and any incoming warm airmass work on them. That means being on and off the glaciated section early, and it means treating the return leg as a different, worse crossing than the outbound one, on ground you have already loaded once. Rain, a warm front, prolonged temperatures above freezing, or a foehn wind should all move you toward not going at all. Committing to a route that only works if you are back before the warmth is a plan with no margin, and glaciers punish plans with no margin.
8. Spacing and travel order are decided before the ground gets serious, not during. Put the person with the best information and the clearest picture of the route in front, and make sure the others can see them. Only one machine loads a suspect section at a time, and the rest wait on ground that has been assessed, with room and time to stop short if the leader disappears. Nobody bunches up. Nobody stops in the middle of a crossing to take a photograph. Stopping points are agreed in advance and are on ground you have reason to trust: rock, moraine, the far side of the field, a line already crossed. Everyone knows before the first machine moves where the group is going to regroup.
9. Understand that a machine changes which bridges hold. A bridge is a span carrying a load, and the total mass matters more than how you spread it. Your machine and you together weigh several times what a person on skis weighs, spread over an area that is still tiny compared to the width of a crevasse. That means bridges that carry a probing, roped mountaineer without complaint can fail under a sled. Speed complicates it further: a fast machine loads any given point briefly and can cross ground that would drop it if it stopped, while a stationary machine, a turn, or a track spinning in place is close to the worst load you can apply. All of that means route selection on a machine has to be more conservative than the route a walking party would accept, not less, even though the machine feels like it makes the ground trivial.

Common Mistakes

Following a saved GPS track as though it were a guarantee. Fix: treat a track as one input, navigate by the ground as well, and confirm the track's age and the conditions it was recorded in before you trust any part of it.

• Probing at an angle. Fix: probe vertically every time, because an angled probe that skims a wall or the underside of a bridge feels exactly like solid ground.
• Continuing to ride when the light goes flat. Fix: make flat light a stop condition. Wait for the light, or turn around, because in flat light every surface clue you were relying on has gone.
• Taking the direct line because the party is behind schedule. Fix: build the longer safe line into the plan and into the timings, and treat schedule pressure as a reason to shorten the objective rather than the route.
• Everyone riding the same marginal bridge one after another because the first machine got across. Fix: recognise that repeated loading cracks a bridge internally, and that if a bridge is marginal enough to be discussed it is too marginal to be crossed five times.

The Drill

On safe, flat, non-glaciated snow, dig a trench about a metre wide and deep enough that you cannot feel its floor with a short probe, then roof it with snow blocks and cover it until the surface looks continuous. Walk away, come back, and probe across it with an avalanche probe, vertically, at fifty centimetre intervals, in gloves. Mark on the surface where you believe each edge of the void is. The standard is that you locate both edges within thirty centimetres, and that you can describe out loud what the change in resistance felt like. Then repeat with the probe deliberately angled and notice how easily the void disappears. Do this until the feel of a void is something you recognise rather than something you are hoping to recognise.

← Full snowmobile curriculum