OpenSnow launches PEAKS 2.0 machine learning weather model

OpenSnow has introduced PEAKS 2.0, an updated in-house machine learning weather forecast system designed to provide hyper-local accuracy for mountain weather and snow reports. Unlike conventional models that simulate atmospheric changes based on current states, PEAKS 2.0 utilizes deep learning and actual data from thousands of weather stations to learn from real-world ground conditions.

The technology is designed to address the complexities of mountain terrain, such as the interplay between peaks, valleys, and elevation, which traditional models often overlook. The system predicts precipitation, temperature, and wind simultaneously, offering a resolution of 1.5 miles with hyper-local adjustments down to 50 feet.

Testing indicates the platform outperforms other models, including the National Weather Service and Google’s WeatherNext, in accuracy across various variables. The service is available in the continental United States and select international locations, aiming to assist skiers, snowboarders, and backcountry travelers in making safer decisions.