At Sknowed.com, we believe users should understand where weather information comes from and what its limitations are. Sknowed presents numerical weather-model data in a format designed for skiers and snowboarders. We do not employ meteorologists or issue official weather forecasts.
This guide explains the data currently displayed by Sknowed, how we present it and how to interpret it when planning a ski trip.
1. Where Our Weather Data Comes From
Sknowed currently obtains forecast data through Open-Meteo, using the Global Forecast System (GFS) operated by the United States National Oceanic and Atmospheric Administration.
-
Weather model: Global Forecast System (GFS)
-
Model resolution: Approximately 0.25 degrees, or about 25–28 kilometres
-
Forecast period: Today plus the following 15 days
-
Model updates: GFS produces new model runs four times daily: 00Z, 06Z, 12Z and 18Z
-
Sknowed updates: Sknowed retrieves updated model data several times each day after new runs become available
-
Data provider: Open-Meteo processes the original model output and supplies it to Sknowed through its API
Because the GFS grid is approximately 25 kilometres wide, neighbouring ski resorts may receive similar or identical values—particularly when they fall within the same model grid area. The model cannot fully represent every mountain, valley, ridge or local weather effect within each grid cell.
2. Elevation-Based Forecasts
Sknowed displays forecasts for different elevations at each ski resort, normally representing the bottom, middle and top of the ski area.
The temperature and other elevation-dependent values are supplied through Open-Meteo for the requested coordinates and elevations. Sknowed does not independently calculate temperature lapse rates, perform proprietary elevation downscaling or modify the data to account for individual slopes and terrain features.
The displayed elevation forecasts can help users compare likely conditions across the mountain, but they should not be interpreted as measurements taken at those exact locations.
3. Snowfall and Other Derived Variables
Not every value returned by a weather API is necessarily a variable published directly by the original weather model. Some variables may be calculated or classified by Open-Meteo from the model’s underlying atmospheric data.
Sknowed currently displays the data received from Open-Meteo. We do not independently:
-
convert precipitation into snowfall;
-
apply our own snow-to-liquid ratios;
-
determine whether precipitation will fall as rain or snow;
-
apply historical bias correction;
-
adjust forecasts for slope direction or local terrain; or
-
create a consensus from several forecasting models.
Where snowfall is shown, it should be understood as model-based forecast data supplied through Open-Meteo—not measured snowfall, settled snow depth or a guarantee of conditions at the resort.
4. AI-Generated Model Summaries
Some Sknowed pages include a written outlook generated from the numerical data displayed on the page. The purpose of this summary is to make the model output easier to understand.
The summary may describe expected snowfall, rain, temperature, wind or freezing-level trends. It does not produce a separate weather forecast, alter the underlying model data or represent analysis by a qualified meteorologist. It is an automated plain-language summary of the supplied model output.
Users should refer to the charts and numerical values as the primary source and report any apparent inconsistency between the written summary and the displayed data.
5. How to Read Our Charts
Snowfall
Snowfall values represent forecast accumulation during the stated period. They do not represent existing snow depth or necessarily show how much snow will remain on the ground. Wind, temperature, compaction, melting and local terrain can substantially affect actual accumulation.
Rain and precipitation
Precipitation represents the total water expected from rain, snow or other precipitation. Rain and snowfall values show how that precipitation has been classified in the data supplied to Sknowed.
Temperature
Maximum and minimum temperatures are modelled values for the selected location and elevation. Actual temperatures may differ, especially around ridges, sheltered valleys and exposed slopes.
Freezing level
The freezing level is the approximate altitude at which the air temperature reaches 0°C. It is useful when assessing the possibility of rain or snow at different resort elevations, but it is not always the exact rain–snow boundary.
Wind speed and gusts
Wind forecasts can help identify potentially exposed or uncomfortable conditions. Lift operators make their own operational decisions, so Sknowed cannot predict whether a particular lift will open, close or be placed on hold.
Weather symbols and probabilities
Weather symbols provide a simplified representation of the model data. Precipitation probability expresses uncertainty and should be considered alongside the expected precipitation amount.
6. Why Forecasts Change
Numerical weather forecasts are estimates of how the atmosphere may develop. Each new model run incorporates updated information and may produce a different result.
Forecast uncertainty generally increases further into the future. A forecast for tomorrow will usually be more reliable than one for two weeks from now. Long-range values are best used to identify broad trends rather than precise snowfall totals or exact timing.
Mountain forecasts are particularly challenging because conditions can change over short distances. Weather may differ significantly between the base and summit, between neighbouring valleys or on opposite sides of the same mountain.
7. Historical and Future Data
Where available, Sknowed’s extended table shows:
-
the previous 15 days;
-
today; and
-
the following 15 forecast days.
The previous days provide recent context, while the forward-looking portion represents the current 16-day model outlook. Historical columns should not be mistaken for part of the future forecast.
8. Verification and Limitations
Sknowed does not currently claim to have completed formal forecast-skill studies or published backtesting against resort observations. We therefore do not publish unsupported accuracy percentages, bias measurements or Mean Absolute Error figures.
We hope to improve our verification methods as suitable observation data becomes available. Any future performance claims should clearly identify the model, location, forecast lead time, observation source and evaluation method.
For important or safety-related decisions, users should consult official weather services, resort announcements, avalanche information and local authorities. Sknowed should be treated as a trip-planning resource rather than an emergency-warning service.
9. Future Model Options
Sknowed currently relies primarily on GFS data. We are investigating whether additional models could improve the service in particular countries.
For Japan, this may include the Japan Meteorological Agency’s higher-resolution Meso-Scale Model (MSM) for short-range guidance, followed by GFS for the longer-range outlook. If multiple models are introduced, Sknowed will explain which model supplies each part of the displayed forecast.
10. Corrections and Feedback
We want the information on Sknowed to be presented accurately and transparently. If you notice an incorrect resort elevation, inconsistent written summary, missing data or another problem, please contact us.
We welcome corrections, recommendations and suggestions for improving our forecasts and charts.
Conclusion
Sknowed organizes third-party numerical weather-model data into resort-specific charts and summaries for skiers and snowboarders. We aim to make that information practical and easy to understand without overstating its accuracy or presenting it as an official forecast.
Weather models are useful planning tools, but they are not guarantees. Check the latest model run, compare elevations, pay attention to forecast uncertainty and consult official local information before making safety-critical decisions.
