Where the real forecasts live

Every Alpine country's national met service publishes a dedicated mountain product — these, not generic phone apps, are the sources worth planning on:

  • Météo-France "Météo Montagne" — per-massif forecasts for 7 massifs, avalanche bulletins (BRA) and snow depths.
  • MeteoSwiss — the official Swiss forecasts, plus a live natural-hazard warning map.
  • GeoSphere Austria — the merger of ZAMG and the Geological Survey since 1 January 2023 (old zamg.ac.at Bergwetter links redirect there). Its Innsbruck meteorologists also write the Alpenverein's 7-day Alps mountain-weather forecast.
  • NOAA/NWS point forecasts (US) — click any coordinate on the map and you get an official forecast for that point at its elevation.

Sources: Météo-France montagne; MeteoSwiss — hazards map; GeoSphere Austria; GeoSphere Wetterportal; Alpenvereinswetter; NWS point forecast (example).

The five things to pull out of any forecast

A mountain forecast is read for altitude-specific elements, not the valley icon. The UK Met Office mountain product — 10 areas, issued daily — is the clearest model of what a complete mountain forecast contains, and it teaches you what to hunt for in any provider:

  1. The freezing level — the altitude where rain becomes snow and wet rock becomes ice.
  2. Wind speed, direction and maximum gusts per altitude band — the Met Office gives them at valley level, 300 m, 600 m and 900 m; ridge wind is what decides your day, not valley wind.
  3. Temperature and "feels like" at altitude — again per band, not one number for the whole mountain.
  4. Visibility and the probability of cloud-free hilltops — the Met Office prints an actual percentage.
  5. Flagged mountain hazards — including thunderstorm risk.

If your usual source cannot answer those five questions for the elevation you are going to, it is a valley forecast wearing a mountain photo.

Sources: Met Office — mountain forecasts; Met Office — Lake District example.

The lapse rate: do the subtraction before you pack

The rule of thumb has an official pedigree: the FAA's Pilot's Handbook gives the standard temperature lapse rate as approximately 3.5°F or 2°C per thousand feet, and the −6.5°C per 1,000 m gradient is the defining tropospheric value of the U.S. Standard Atmosphere 1976.

Applied to a trekking day: a valley forecast of 20°C means roughly 7°C at a ridge 2,000 m higher — before wind. The pleasant number on the phone and the conditions at the col are separated by simple arithmetic that most people never do.

Sources: FAA — Pilot's Handbook, ch. 4 (PDF); U.S. Standard Atmosphere 1976 (NOAA PDF).

Afternoon storms, pre-dawn starts

Summer mountain thunderstorms follow an afternoon convective pattern, and the official park guidance is built around it. Grand Teton calls afternoon thundershowers "common" in July and August. Rocky Mountain's Longs Peak page: "In June, July and August, be aware of afternoon thunderstorms and lightning", and be "prepared to turn back during sudden, drastic weather changes". Great Sand Dunes tells summer visitors to plan their exposure for early morning or evening to avoid the storms.

An honesty note our sources force on us: the popular slogan "summit by noon" appears in no verified official source. What the official pages actually say is subtler and better — start early, know that the risk builds after midday, and predecide that you will turn around when the weather turns. The slogan is a fixed deadline; the guidance is a pattern plus a commitment.

Sources: NPS Grand Teton — weather; NPS Rocky Mountain — Longs Peak; NPS Great Sand Dunes — safety.

Lightning above treeline: the current doctrine

The NWS states the baseline in capital letters: "NO PLACE outside is safe when thunderstorms are in the area!!" And the trigger is sound, not sight: "If you hear thunder, lightning is close enough to strike you."

Above treeline, where no safe shelter exists, the NWS last-resort playbook is risk reduction, not protection:

  • "Immediately get off elevated areas such as hills, mountain ridges or peaks."
  • "Never lie flat on the ground."
  • "Never shelter under an isolated tree."
  • "Never use a cliff or rocky overhang for shelter."
  • Get away from water, and "stay away from objects that conduct electricity."

Two pieces of folklore fail against current guidance. The lightning crouch is no longer endorsed as protective — the NWS myths page says it outright: "Crouching doesn't make you any safer outdoors. Run to a substantial building or hard topped vehicle." And the old "spread the group out" advice no longer appears on any current NWS page, so we do not repeat it as official doctrine.

The stakes, from NWS data: US lightning deaths averaged 27 per year over 2009–2018 (down from 43 per year across 1989–2018), with roughly 243–270 injuries a year; about 10% of people struck are killed. The NWS analysis of 418 deaths from 2006–2019 found almost two thirds happened to people enjoying outdoor leisure activities, and more than 70% of deaths fell in June–August — exactly the trekking season.

Sources: NWS — lightning safety; NWS — lightning safety tips; NWS — lightning myths; NWS — lightning odds; NWS — fatality analysis 2006–2019 (PDF).

Cold kills above freezing

Wind chill, in the NWS definition, "is a 'feels like' temperature based on the rate of heat loss from exposed skin caused by the effects of wind and cold. As the wind increases, human and animal bodies are cooled at a faster rate." The official chart lives on the NWS cold-safety page; the mechanism is the reason ridge-wind numbers matter as much as the temperature line.

The half of the story trekkers underestimate is summer hypothermia. The NWS is explicit that hypothermia "can also occur at cool temperatures (above 40°F) if a person become chilled from rain, sweat, or submersion in cold water." That is well above freezing — a wet, windy summer afternoon at altitude is enough. Cold injuries are a medical matter: this is general safety information, not medical advice, and in an emergency you call the local emergency number first.

Sources: NWS — wind chill chart; NWS La Crosse — winter safety.

Whiteout: when the phone dies, the plan must not

Mountaineering Scotland — the national mountaineering body — describes how fast the picture changes: mountain weather "can very quickly change to one of poor visibility, demanding good navigation techniques", and falling snow "will blot out much of the detail in the landscape." Its verdict on single-device navigation is blunt: "No system is 100% fail safe, so having a separate system to navigate with if the phone or GPS were to fail is a must, for example a map and compass."

This is not hypothetical caution. The same body reports: "Mountain rescue teams respond to many call-outs because of GPS/smartphone failure, inappropriate route choice or because people are unable to use the technology effectively." The Swiss numbers put a size on the trend from another angle: in 2025, 38% of the nearly 4,000 people rescued from mountain-sport emergencies in Switzerland were unhurt — up from 29% in 2020. Getting stuck, not getting hurt, is the fastest-growing reason for a rescue call.

The counter-skills are old ones: navigate on large terrain features, use timing and pace-counting, and prepare bearings for the critical legs before you need them — a whiteout is a bad classroom for compass work.

Sources: Mountaineering Scotland — navigation; winter navigation; when navigation goes wrong; SAC-CAS — Bergnotfallstatistik 2025.

Why the app misleads in terrain: 13 km versus 2.8 km

The physics, from the German national met service (DWD): its operational global model, ICON, has "an effective mesh size of about 13 km", and in DWD's own words, "While large mountain ranges can be represented reasonably well at a mesh size of 13 km, capturing individual mountain ranges or valleys requires even finer meshes." That is exactly why national services run limited-area models at around 2.8 km over their own terrain.

The practical conclusion follows from the arithmetic, not from any app's marketing: a forecast interpolated from a ~13 km global grid cannot resolve your individual ridge, your valley, or the convection cell building over either. Between two forecasts for the same summit, the one downstream of a national ~2.8 km model — the sources in the first section — deserves your trust; the one interpolating a global model deserves your skepticism, however precise its hour-by-hour numbers look.

Read the right forecast, extract the five numbers, subtract the lapse rate, respect the afternoon, and carry a navigation system that doesn't run on a battery. That is the whole method.

Source: DWD — ICON model description.

Sources