How to Assess Snowpack Stability: A Data-Driven Guide

01What Is Snowpack Stability? Reading the Layered History Book

Snowpack stability is how likely the snow on a slope is to avalanche. You assess it by combining three data streams: the official avalanche bulletin, the recent weather (the loading history) and what you see and feel in the field — buried weak layers, whumpfing, shooting cracks, recent avalanches. No single pit or test proves a slope is safe.

Stability is a judgement built from converging evidence. Every backcountry slope you ski is a vertical archive. The snowpack is not a single homogeneous block — it is a stack of layers, each one the frozen record of a single weather event: a storm, a windy afternoon, a clear cold night, a thaw, a refreeze. By mid-winter the snowpack on a north-facing Alpine slope may hold 8 to 15 distinct layers within a depth of 120–200 cm.

Avalanches happen when a cohesive slab (a bonded layer of snow) sits on top of a weak layer that fails, all of it resting above a smooth bed surface. Three ingredients must coexist: a slab, a weak layer beneath it, and a slope steep enough to slide — almost all slab avalanches release on slopes between 30° and 45°, with the statistical peak around 38°. Below ~30° dry slabs rarely release; above ~50° snow tends to sluff continuously rather than accumulate into dangerous slabs.

The entire discipline of snowpack stability assessment is really one question asked in many forms: is there a weak layer buried in this history book, how reactive is it, and how widely will a fracture in it spread? The rest of this guide breaks that question into data you can actually observe and read.

02Snow Metamorphism: The 10°C/m Threshold That Decides Everything

Once snow lands, the crystals immediately begin to change shape — a process called metamorphism. Which direction they change is governed almost entirely by the temperature gradient: how fast temperature changes with depth through the snowpack.

The single most important number in snow science is this: *a gradient steeper than ~10°C per metre (equivalently 1°C per 10 cm) drives kinetic growth — faceting — which weakens the snow. A gradient gentler than that drives equilibrium* (rounding) metamorphism, which strengthens it.

The physics: the ground beneath the snow stays near 0°C all winter, while the snow surface on a cold clear night can drop to −15°C or colder. The temperature difference drives water vapour upward from the warmer snow near the ground to the colder layers above. Under a weak gradient, vapour deposits gently and grains round off, bond, and sinter — this is good, strong snow. Under a strong gradient, vapour moves so fast that grains grow new angular, stepped, cup-shaped faces with almost no bonds between them. These facets are the seed of nearly every persistent avalanche problem.

The cruel arithmetic of this is that a thin snowpack is a dangerous one. A 60 cm cold snowpack packs the whole 0°C-to-−15°C difference into 60 cm — a gradient of 25°C/m, deep in faceting territory. The same air over a 200 cm snowpack gives only 7.5°C/m — safely rounding. This is why early-season shallow snow, and shallow rocky zones within an otherwise deep pack, are so often the trigger points.

Same air temperature, opposite outcome: the 200 cm pack sits at 7.5°C/m and strengthens, while the 60 cm pack hits 25°C/m — well above the 10°C/m faceting threshold — and rots into depth hoar. Thin snow is weak snow.

03What Is a Weak Layer? The Three Persistent Types to Know

A weak layer is a band of fragile, poorly-bonded snow crystals buried inside the snowpack — most often surface hoar, faceted grains, or depth hoar. When a fracture starts in it, the stronger, cohesive slab resting on top can slide off as a slab avalanche. Finding and characterising these layers is the core of stability assessment.

Weak layers come in two families. Non-persistent problems (new storm snow, wet snow) stabilise within hours to a couple of days. Persistent weak layers can stay reactive for weeks or even the entire season, and they are responsible for the deadliest, least-predictable avalanches. There are three you must be able to name and recognise.

1. Surface hoar — the frozen equivalent of dew. On clear, cold, calm, humid nights, feathery crystals grow on top of the snow surface, sometimes 5–40 mm tall and glittering in the sun. Beautiful, and deadly: when buried by the next storm, surface hoar becomes a near-frictionless, sheet-like weak layer that fractures and propagates across entire bowls, even into low-angle terrain. Once buried it often stays reactive for weeks, sometimes for months (avalanche.org).

2. Faceted snow (near-surface and mid-pack facets) — angular grains formed by the strong gradients described above, often around buried crusts where vapour pools. Sugar-like, non-cohesive.

3. Depth hoar / basal facets — the largest faceted grains, cup-shaped and striated, up to 5–10 mm, formed at the very base of shallow, cold snowpacks under gradients exceeding 10°C/m sustained for weeks. Depth hoar collapses with an audible whumpf and produces full-depth, often unsurvivable slides. It is the classic continental-snowpack killer but appears in the Alps in cold, snow-starved early seasons and in shallow, shaded, rocky terrain.

Where the Alps fit — and how to track a buried layer. Snowpacks are broadly classed as maritime (deep, warm, rarely faceted), continental (thin, cold, chronically faceted — depth hoar's home turf), or intermountain (transitional) in between. The Alps do not fit one class: an SLF study of 128 Swiss weather stations (Schweizer et al., ISSW 2024) classed 45% of the station-winters as transitional, 34% as maritime (dominant on the northern slopes) and 21% as continental (notably in parts of the Valais and the Engadine). So persistent problems can come and go with cold snaps on the wetter northern side, yet linger for much of the winter in the drier inner Alps. Once a surface-hoar or facet layer is buried, follow it: check how its reactivity trends in successive bulletins over the days and weeks it stays dangerous, rather than assuming a few sunny days have healed it.

What our own records show. Snow Trace's avalanche dataset holds 3,449 French avalanche observations (data-avalanche.org, February 2016 to February 2026). Of the 3,306 that record an aspect, 1,738 (about 53%) faced north-west, north or north-east. That fits the physics above: shaded slopes stay colder, so facets and buried surface hoar last longer there. But the records carry no avalanche type, aspect is not a cause, and shaded slopes may simply be skied more. Read it as a reason to check the bulletin's critical aspects, not as a rule.

The defining property of all three: they do not heal quickly. A storm slab forgives you in 48 hours. A buried surface-hoar or depth-hoar layer can punish a decision made a month later, which is why bulletins flag a persistent weak layers problem (in North America: Persistent Slab or Deep Persistent Slab) long after the last snowfall.

04Slab Avalanche Types: Wind, Storm & Persistent Slabs

A weak layer is only dangerous if there is a cohesive slab sitting on it. Three slab problems dominate decision-making, and each has a different signature.

Storm slab — fresh, bonded new snow overloading a buried interface. Highest danger during and 24–48 hours after a storm. Sensitive to snowfall rate: loading faster than ~2–3 cm/hour, or a total of >30 cm of new snow, sharply raises reactivity. Generally stabilises fast. Judge how much new snow actually fell by weight, not just depth — a dense, high-water-content storm loads a buried weak layer far faster than the same centimetres of light, low-density powder.

Wind slab — wind transports snow up to 5–10× faster than it falls, stripping windward slopes and depositing dense, chalky, hollow-sounding slabs on leeward aspects and in cross-loaded gullies. Wind slabs form even on bluebird days with no new snow, are often localised, and are one of the most frequent skier-triggered problems in the Alps. Look for smooth, pillowy, matte-textured snow and cornices pointing to the lee aspect.

Persistent slab — a slab over one of the persistent weak layers above. The dangerous trait is low spatial predictability and remote triggering: you can trigger it from flat terrain below, or from a thin spot on the slope, and the fracture can run hundreds of metres. These demand the widest safety margin and the most conservative terrain choices.

05Reading the Avalanche Bulletin: The EAWS 1–5 Scale

In the Alps, your most important data source is the daily regional avalanche bulletin, issued on the European Avalanche Warning Services (EAWS) five-level scale. Crucially, the scale is not linear — danger and the number of avalanche-prone slopes roughly double with each step up.

The counter-intuitive but vital statistic: the **majority of avalanche fatalities occur at levels 2 (Moderate) and 3 (Considerable)** — not at 4 or 5. At high levels people simply stay home; at moderate levels the hazard is patchy, tempting, and easy to underestimate. Treat the level as a starting point, never a green light.

Don't read only the number. Every bulletin specifies which avalanche problem is active, the **critical aspects and elevations* (shown on a rose/clock diagram), and a trend. A 'Considerable, persistent weak layers, north through east above 2200 m' tells you exactly* which slopes to avoid — far more actionable than the headline digit alone. If your bulletin is Météo-France's BERA (bulletin d'estimation du risque d'avalanche), our field-by-field guide to reading the BERA breaks down the rose and problem icons.

The scale is exponential, not linear: avalanche-prone terrain roughly doubles each step, so Level 5 carries about 16x the hazard of Level 1. Yet roughly 80% of fatalities strike at Levels 2-3 (Moderate & Considerable), where the hazard looks modest but the terrain is tempting and people keep skiing. The danger number is a starting point - never a green light.

06How to Test Snowpack Stability: ECT, Compression Test and Rutschblock

To test snowpack stability, dig a pit and isolate a column of snow, then load it in steps to see whether a buried weak layer fails — and, just as importantly, whether that fracture propagates. The three standard field tests are the Extended Column Test (ECT), the Compression Test (CT) and the Rutschblock (RB). The bulletin is regional; your slope is local, and these tests let you sample the layering yourself. They are most useful for finding and characterising weak layers — a clean, sudden failure is meaningful evidence of instability — but a 'stable' result on one test pit does not prove a slope is safe. Treat them as one data point among many, never as a clearance. Reading a pit well takes practice: it is exactly what a hands-on avalanche course teaches.

One pit is not the slope. Snowpack structure varies across a bowl, so where you dig matters as much as how. Pick a small, safe, representative test slope — same aspect, elevation and wind exposure as your objective, never the start zone itself — and read one result as a single sample, not a verdict. Professionals dig several pits and weight the pattern, because a stable column two metres away can hide a reactive one just uphill.

Compression Test (CT) — Isolate a 30 × 30 cm column. Tap from the wrist (10 taps), then elbow (10), then shoulder (10), recording when and how a layer fails:

• CT 1–10 (Easy): very weak, alarming
• CT 11–20 (Moderate): suspect
• CT 21–30 (Hard): stronger, not safe by itself

The fracture character matters more than the number: a sudden, clean 'pop' (SP/SC) that slides as a block indicates a propagating weak layer and is a strong red flag. A resistant, rough, non-planar break is less concerning.

Rutschblock Test (RB) — A larger 2 m × 1.5 m block isolated on three sides, loaded by a skier in progressive steps (RB1 = fails while isolating, up to RB7 = won't fail even when jumped). It samples a more realistic, skier-sized area:

• RB 1–3: poor stability — fractures under low load
• RB 4–5: fair — fails under heavier load
• RB 6–7: good — hard to trigger

Extended Column Test (ECT) specifically tests propagation: an 'ECTP' (the fracture runs across the full 90 cm column on the tap that initiates it or the next one) is one of the clearest warnings that a fracture will spread. The golden rule for all tests: they can confirm a slope is dangerous, but they can never confirm it is safe.

Forecasters compress all of this — snowpack structure, test results and observations — into a stability class from very poor through poor and fair to good (the four classes of the EAWS Matrix behind the danger level). Map your own field evidence onto that scale: a clean, propagating ECTP is a very poor to poor signal, while a stubborn, non-propagating result is merely fair — never a green light on its own.

ECTP means the crack propagated across the full column — the single clearest test warning that a slope can release. ECTN (a layer that initiates but does not propagate) is more ambiguous, and ECTX (no fracture in 30 taps) is never a clearance: it is one stable data point, not proof the slope is safe.

07The Daily Go / No-Go Decision Framework

Stability assessment becomes useful only when it converts into a decision. Combine four data streams in order, treating any single strong red flag as a veto.

Step 1 — The Bulletin (the night before & morning). Note the danger level, the active avalanche problem(s), and the critical aspect/elevation rose. This sets your terrain budget. Snow Trace's tourist-office pages link the official bulletin for their area, and our destination guides for Chamonix, Zermatt, Val d'Isère and Verbier each point to the bulletin that covers them.

Step 2 — Recent weather (the loading history). The biggest stability drivers are:

• New snow: > 30 cm in 24 h, or > 20 cm on a buried weak layer, = elevated storm-slab danger.
• Wind: moderate wind (~25–40 km/h) over loose snow builds slabs most efficiently; expect fresh wind slabs on lee aspects.
• Rapid warming / first sun: spring warming or a +5°C swing can trigger wet-loose and wet-slab cycles within hours (see the melt-freeze corn cycle).
• Rain on snow: an immediate, severe red flag.

Step 3 — The 'Big Five' observable red flags in the field. Any one means scale back:

1. Recent avalanches on similar aspects/elevations.
2. Whumpfing — collapsing sounds (a weak layer failing under you).
3. Shooting cracks radiating from your skis.
4. Heavy recent loading (snow or wind).
5. Rapid warming / wet snow (rollerballs, pinwheels, ski-pole punch-through).

Step 4 — Terrain choice (the only variable you fully control). You cannot change the snowpack; you can change the slope you stand on — so plan your route and turnaround time around the safe aspects and elevations before you leave the car. Apply the Reduction Method heuristic — at higher danger levels, cap your maximum slope angle:

Measure slope angle with an inclinometer or your phone — eyeballing is unreliable, and an error of 5–10° is easy to make. Remember: a gentle slope is only safe if nothing steep looms above it (connected terrain and runout zones count).

Snow Trace's route graph holds 6,345 ski-touring routes, and 23 massifs have their own regional profile. On the map, the avalanche bulletin layer colours each micro-region with its maximum danger level only: no avalanche problem, aspect or elevation. Use that colour as a pointer to the full bulletin, never as a substitute for it.

If the bulletin, the weather, and the field all agree the conditions are favourable for your chosen terrain — go, with spacing, one-at-a-time on suspect slopes, on a reliable touring setup you trust, and with rescue gear on (plus the drilled skills to use it fast). If they disagree, the conservative read wins. The mountain will be there next week. To learn pit work and terrain choice in the field, find an avalanche course or a certified pro.

08Sources & Further Reading

Sources & further reading. This article reflects the consensus of the major avalanche-safety organisations and the standard references. Always defer to your local daily avalanche bulletin and hands-on training over any single article:

• **Avalanche.org** — US National Avalanche Center: forecasts and free avalanche education
• **EAWS** — European Avalanche Warning Services: the standard danger scale and avalanche problems
• **SLF** — Swiss Institute for Snow and Avalanche Research
• Bruce Tremper, "Staying Alive in Avalanche Terrain" — the standard recreational reference
• McClung & Schaerer, "The Avalanche Handbook" — the standard technical reference