How to read an avalanche forecast properly

An avalanche forecast is a dense, structured document written by professionals who have spent the day in the snowpack. Most recreational users extract one number from it and discard the rest, which is a bit like reading a weather forecast and noting only whether it says "changeable".

The danger rating is a summary, not an instruction

The five-level scale runs from low through moderate, considerable and high to extreme. It is useful for a rough sense of the day, and it has two well-documented problems.

First, the scale is not linear. Each step represents a substantially larger increase in hazard than the step before it.

Second, and more importantly: a large share of avalanche fatalities occur at considerable, the middle of the scale. Not at high, and not at extreme. The reason is that at high and extreme people stay home, whereas considerable is an ambiguous rating on a day that often looks perfectly pleasant, and people go out and make judgement calls.

So the rating tells you very little about what to do. The rest of the forecast tells you a great deal.

The avalanche problems

This is the part that matters most. Forecasts identify specific problem types, each with a location, a likelihood and a size.

The type tells you how it behaves. A storm slab is reactive now and settles in days. A wind slab sits on lee aspects and stabilises fairly quickly. A persistent slab can stay dangerous for weeks or months and propagates widely. A wet loose problem develops through a warm day. A cornice problem is about ridges. Each requires a different response, and treating them identically is the most common analytical error.

The location — expressed as aspects and elevation bands, usually on a rose diagram — is an instruction. If the rose shows the problem on north through east above treeline, that is precisely the terrain to avoid, and it also tells you where you can go.

The likelihood and size together give consequence. A very likely small avalanche is a different day from an unlikely but very large one, and the second is often the more serious because size determines burial depth and trauma.

The discussion or summary text

Frequently the most valuable section and frequently unread. This is where forecasters write their reasoning, their observations, and — crucially — their uncertainty.

When a forecaster writes that they are unsure how a layer is behaving, or that they have limited observations from a particular area, that uncertainty should propagate directly into your margin. Professional forecasters are careful about hedging, and when they hedge explicitly it means something.

The weather and snowpack sections

Recent weather explains why the problems exist. New snow amounts and rates. Wind speed and direction, which tells you which aspects were loaded. Temperature trends, freezing levels, and whether there was a good overnight freeze.

The snowpack summary tracks the season's structure — when weak layers formed, how deep they are, how they have been behaving.

Read these and the problems stop being arbitrary and start being predictable.

Read the trend, not the snapshot

Read the last several days of forecasts, not only today's. Is the danger rising or falling? Is a persistent problem being described as improving, or as stubborn?

A moderate rating on the way down from high is a different situation from a moderate rating that has been stable for a week, and a single day's document cannot show you that.

Read the observations

Most forecast centres publish field observations from professionals and from the public: avalanche activity, whumpfing, test results, weather. This is raw data from the terrain, often more current than the forecast itself.

Recent avalanche activity on a given aspect and elevation is the strongest possible evidence about that aspect and elevation. Look for it before every trip.

Turning it into a plan

The forecast should produce three concrete things before you leave the house:

A terrain limit — the maximum slope angle you will travel on or beneath.

A list of aspects and elevations to avoid, taken directly from the rose.

An alternative plan for lower-consequence terrain if conditions on the ground differ from the forecast.

Write them down. A forecast that has not changed your route has not been used.

The limits of a forecast

It is a regional product covering a large area with enormous local variation. It is written the previous evening or early morning and conditions change. It cannot know about the specific slope you are standing on.

So it is a starting point, not an answer. On the ground you are checking whether what you see matches what was predicted — and when it does not, the ground wins, every time.