WindBorne Systems has launched its latest AI weather forecasting model, WeatherMesh 6, which promises enhanced accuracy and frequency in predictions. This new version generates forecasts every hour, a significant improvement over the six-hour interval typical of traditional systems. The model boasts a resolution of 3 km in Europe and the continental US, where data quality is optimized.
Founded in 2019 by Stanford students, WindBorne initially focused on developing a superior weather balloon for data collection. However, the introduction of deep learning models for weather forecasting in 2022 prompted the team to create their own predictive model. WindBorne claims WeatherMesh 6 provides forecasts that are as precise five days in advance as traditional methods are just one day ahead, particularly excelling in surface temperature predictions.
The company currently operates approximately 400 balloons globally, collecting crucial sensor data from 15 launch sites. This innovative approach positions WindBorne as a competitive player against the European Centre for Medium-Range Weather Forecasting (ECMWF), recognized as a leader in weather prediction accuracy.