# Can avalanches in Hokkaido be predicted? Forecasts give a regional probability

Forecasts give a regional probability; research says no single avalanche can yet be predicted in time and space. So ask: any reason to skip this slope today?

- URL: https://www.ykhokkaido.com/en/blog/bc-prep-02-can-avalanches-be-predicted
- Published: 2026-11-10
- Section: Backcountry
- Series: Backcountry Prep (part 2)
- Tags: Avalanche safety

> This post is the author's personal experience and notes; it is not a safety guarantee. For backcountry outside the managed area, take a formal avalanche course and go with a qualified guide. snowwind.org is SnowWind's sister brand (backcountry).

---

This post answers one question: what an avalanche forecast can tell you, and what it can't.

This post is the author's personal experience and notes; it is not a safety guarantee. For backcountry outside the managed area, take a formal avalanche course and go with a qualified guide. snowwind.org is SnowWind's sister brand (backcountry).

This post is for you if you:

- are considering backcountry riding in Hokkaido and want to know how far a forecast can be trusted
- have seen avalanche danger ratings and aren't sure how to use them
- want to know why "Is it safe today?" has no standard answer

## What a forecast is saying

In 2003, Jürg Schweizer and colleagues wrote a review paper on avalanche formation. The first part of the abstract says that in many mountain regions with a seasonal snow cover, avalanche forecasting services publish the probability of avalanches for a given region to warn the public.

Then comes this sentence:

<blockquote class="jq">
<p>It is not yet possible to predict a single avalanche event in time and space.</p>
<cite>Schweizer et al. (2003), rendered from our Chinese summary of the abstract</cite>
</blockquote>

The abstract gives several reasons. The snowpack is a highly variable layered structure, and it sits close to its melting point. The interactions between terrain, snowpack and weather are complex. The authors say the field still has many gaps in its knowledge.

<figure class="jd">
<svg viewBox="0 0 400 240" role="img" aria-labelledby="bp02-slope-title">
<title id="bp02-slope-title">Mountain profile: one probability band over the whole region, a question mark on a single slope</title>
<path d="M20 200 L120 120 L220 90 L380 200" fill="none" stroke="currentColor" stroke-width="2"/>
<rect x="20" y="30" width="360" height="26" fill="none" stroke="var(--c-blue)" stroke-width="3"/>
<text x="30" y="49" fill="currentColor" font-size="15">Forecast: probability for this region</text>
<text x="130" y="140" fill="currentColor" font-size="32">?</text>
<text x="215" y="150" fill="currentColor" font-size="15">This slope: no answer</text>
</svg>
<figcaption>Schematic. The forecast gives the whole region one probability; it gives no answer for a single slope.</figcaption>
</figure>

A 2006 study by Betty Sovilla and colleagues used observations from test sites in Switzerland and Italy, and from avalanches that occurred in Switzerland, to analyse the mass and energy of 18 avalanche events. On average, their mass increased fourfold. The study is about how avalanches pick up snow along their path; this post cites only that one result, to make one point: once an avalanche starts, it can become much bigger than the release you see.

**Key point**: a forecast is a regional probability, not an answer for the slope under your feet.

## The Hokkaido picture

Japan publishes avalanche information. 西西 has written up the format used by the Japan Avalanche Network (JAN), whose bulletins cover Niseko, Rusutsu and the Shiribetsu area:

- **Danger level**: five levels — Low, Moderate, Considerable, High and Extreme — usually with a trend arrow beside it.
- **Avalanche type**: seven types, such as wind slab and wet snow avalanches, marked with the aspects and elevation bands where they are likely.
- **Summary and recommended actions**: written by staff, reviewing whether any avalanches occurred the day before, and recent weather and snowpack conditions.

西西 also points out that JAN's information is an assessment of the current situation on the day, not a multi-day forecast, so you still have to judge from the snow you see yourself. This matches the paper's conclusion above: regional information, not an answer for a single slope.

西西 has also listed a few things to observe. Only the directions are given here; learn the details on a course:

- Terrain: slope angle, hollows, and ridges where cornices may form.
- Weather: temperature changes, the amount and rate of snowfall, wind direction and speed, pressure systems. 西西 says heavy snowfall in a short time significantly increases the risk.
- Snowpack: depth hoar is an unstable snowpack structure, and 西西 specifically warns to watch for it after heavy snowstorms.

Trees are no guarantee either. 西西's notes say that big trees do not mean safety: the presence of trees may only mean that full-depth avalanches happen less often at that spot, which does not rule out surface avalanches. An avalanche on the slope above can also reach the trees.

Inside the managed area and outside it are two different things. SnowWind lessons always take place inside the resort's managed area, in terrain the resort permits and has open that day. To learn about terrain and snow conditions outside the managed area, see our sister brand snowwind.org. Backcountry riding is at your own risk; always go with a qualified local guide.

**Key point**: information like JAN's is a regional assessment for the day; you still have to judge from the snow you see yourself.

## The decisions you can make

A forecast cannot answer "Will this slope slide?", so the question you can ask changes: today, on this slope, do I have a reason not to go? That framing is our editors' summary, not a conclusion of the forecast.

Among 西西's reminders about riding the backcountry as a group, one line deserves to stand on its own:

<blockquote class="jq">
<p>Don't walk into danger just because you feel awkward turning down a partner's idea. Many people have been caught in avalanches because they felt too awkward to refuse a dangerous suggestion.</p>
<cite>西西</cite>
</blockquote>

西西 also lists principles for working as a team, such as letting only one person at a time cross a hazardous area while the others watch from a safe spot. These are only pointers.

Rescue details are another matter. In a Threads post in early November 2026, 西西 wrote that after attending an avalanche course, 西西 heard about many details that are easy to overlook. For example, at first you hold the transceiver (beacon) to your ear and listen, with your eyes on the slope rather than on the screen. Things like this have to be practised on a course; an article cannot teach them.

The decisions you can make fit on one list:

- Before going outside the managed area, take a formal avalanche course.
- Choose a qualified guide, and make sure that guide is willing to say "not today".
- In your group, make sure everyone can say no.
- For gear, see the [Backcountry gear checklist](/en/blog/backcountry-gear-checklist/).

**Key point**: a forecast gives you information; the decision is still yours. And saying no is part of the decision.

## Who this post is not for

- People looking for a standard "do this and you will be safe" procedure: there is no such procedure, and this post does not offer one.
- People hoping to learn avalanche rescue or snowpack assessment from an article: take a formal course.

For why Hokkaido powder is the way it is, see [The Science of Powder, Part 1](/en/blog/powder-science-01-why-hokkaido/). Common questions are answered in the [FAQ](/en/faq/).

<p class="jn">The papers are cited only for conclusions their abstracts state outright, taken from the site's Chinese summaries of the papers; the English wording here is translated from those summaries. The JAN, terrain and weather material comes from 西西's teaching notes (the original does not give a year); the rescue detail comes from 西西's Threads post of 5 November 2026. Threshold figures in these sources, such as slope angles and snowfall rates, have been left out, because they should not be treated as safety limits.</p>

## FAQ

### Are avalanche forecasts accurate?

A forecast publishes the probability for a region, not an answer for a single slope. The abstract of a 2003 review by Schweizer et al. states that single avalanche events cannot yet be predicted in time and space. Use a forecast as a reference, not as proof that a slope is safe.

### Can avalanches be predicted?

Regional probabilities can be published; a single avalanche cannot be predicted. The abstract's reasons: the snowpack is a highly variable layered structure close to its melting point, and terrain, snowpack and weather interact in complex ways.

### Can I go into the backcountry on a storm day?

This post gives no yes-or-no answer. Among the weather factors 西西 has compiled, heavy snowfall in a short time significantly raises the risk. Outside the managed area, take a formal avalanche course and go with a qualified guide; don't let one article decide.

### Do SnowWind lessons go into the backcountry?

Lessons always take place inside the resort's managed area, in terrain the resort permits and has open that day. Backcountry riding is at your own risk; always go with a qualified local guide.

## Sources

- Jürg Schweizer et al. (2003). Snow avalanche formation. https://doi.org/10.1029/2002rg000123
- Betty Sovilla et al. (2006). Field experiments and numerical modeling of mass entrainment in snow avalanches. https://doi.org/10.1029/2005jf000391
- https://www.threads.com/@ryokan_ijuin/post/DfXXBb6GUKo
- https://www.threads.com/@ryokan_ijuin/post/DfXTCwSGYAM
