NivoPeak

The name is a combination of two elements. "Nivo" comes from the Latin word for snow and is the root of nivology, the scientific study of snow and avalanches. "Peak" represents the alpine environments our forecasts target.

Because of this hybrid origin, we embrace two pronunciations: an international English one, and a local Alpine one that honors the Latin root of "Nivo".

All the weather and snowpack data shown in our forecasts is provided by Open-Meteo , an open-source weather API that aggregates open-data models from national meteorological agencies such as MeteoSwiss (Switzerland) and ItaliaMeteo (Italy).

NivoPeak relies on highly specialized, regional weather models to ensure maximum accuracy in the Alps and surrounding ranges. Models like ICON-CH1 cover Central Europe, while ICON-2I covers Southern Europe. Because these models focus their computing power on specific alpine domains, snowpack data for mountains outside these geographical areas (e.g., the Americas or Asia) is currently unavailable.

Not perfectly. While our models have an exceptionally high spatial resolution (down to a ~1.1 km grid), the supercomputer still calculates an average elevation for that grid square. If you are standing on a sharp peak or in a deep, narrow ravine, your actual elevation might differ from the model's "smoothed" topography, which can slightly influence the local snow depth calculation. Always use this data as a trend, not an absolute measurement.

It defines the grid size of the weather model. A higher resolution (like 1.1 km for ICON-CH1) means the model can better 'see' the complex topography of mountains. This results in far more accurate snow forecasts for specific peaks or narrow valleys compared to lower resolution models.

If lines from different agencies (or models) diverge significantly, it means the atmospheric conditions are highly unstable and hard to predict. Different models use different physics and resolutions; a 1.1km model might "see" a localized storm effect that a 2.2km model misses. When forecasts disagree, confidence is low: we recommend trusting the model with the best spatial resolution for that specific region, but preparing for the most conservative scenario.

They represent forecast uncertainty, calculated using 'Ensemble' models. The shaded band shows the range between the 10th (p10) and 90th (p90) percentiles of multiple simulated weather scenarios. A narrow band means high confidence (the scenarios agree); a wide band means the weather is highly unpredictable.

Different models have different forecasting 'horizons'. Ultra-high-resolution models update very frequently but only look 33 hours ahead to maximize short-term accuracy. Other models are designed to look 3 to 5 days into the future.

Supercomputers run these weather models at specific intervals. The status dot (

On time ,

Delayed ,

Error ) monitors the real-time 'health' of the upstream agencies. If a model is delayed, you are seeing the previous run's data.