#None #Envision #Tianji #AI Weather Quest

Envision's Tianji Weather Model Triumphs

In a groundbreaking achievement, Envision, a leader in green technology, announced that its Tianji Weather Foundation Model has clinched first place in the AI Weather Quest competition organized by the European Centre for Medium-Range Weather Forecasts (ECMWF). This competitive showdown saw Tianji outclass nearly 20 top-tier teams from around the globe, securing the top position across all assessed categories including surface average temperature, mean sea level pressure, and precipitation forecasts for the 19 to 25-day prediction window.

Tianji's notable win follows a series of impressive performances, including high rankings in the global WeatherBench 2 assessment, where it rivaled the Google WeatherNext2 model, and a first place finish in deterministic forecasts in the AI Weather Forecast Model Demonstration Program organized by the Chinese National Meteorological Center.

Superior Performance Metrics

The AI Weather Quest evaluates models based on continuous real-time predictions and their sustainability performance under various weather conditions. Beyond the 15-day mid-range forecasting horizon, the competition emphasizes probabilistic forecasts of weekly average temperatures, mean sea-level pressure, and weekly precipitation amounts for days 19–25. This evaluation method assesses temperature and precipitation primarily over land-surface grid cells, while mean sea-level pressure is globally assessed. The inherent focus on predicting weekly conditions and their uncertainties weeks ahead adds a critical layer of complexity to the challenge.

During the competition, participating teams submitted their weekly weather forecasts, which ECMWF matched against actual meteorological data, updating rankings publicly. Envision's Tianji not only demonstrated the ability to deliver accurate short-term forecasts but validated its utility in energy applications. It excelled in recent competitions such as the AI load forecasting contest hosted by State Grid and the business competition organized by China Southern Power Grid.

Robert Zheng, head of the Tianji Weather Foundation Model at Envision, expressed, “Subseasonal forecasting—predicting weather two weeks to two months in advance—represents some of the greatest challenges in meteorology. The longer the forecast horizon, the greater the uncertainty caused by atmospheric changes, demanding models with a robust ability to identify evolving weather trends.”

Pioneering Technology

Leveraging years of practical experience in new energy, Envision has developed the world's first physics-based AI weather model, Tianji, which utilizes high-resolution spatial and temporal data in the petabyte range, assessed through a three-dimensional data network integrating space, air, and ground data to provide reliable forecasts over extended periods. Tianji's use of advanced computing technology facilitates rapid short-term and global predictions with marked improvements in reliability for critical metrics such as wind speed and solar radiation across key regions.

The model's capabilities span a wide range from minute-by-minute forecasting to subseasonal trend analysis, providing comprehensive meteorological insights for real-time monitoring and long-term operational planning. Thanks to its sophisticated algorithmic foundation, Tianji has positioned itself at the forefront of weather prediction technology, setting a formidable precedent for the industry.

Envision's ongoing commitment to innovation ensures that Tianji will continue to evolve, meeting the growing challenges posed by climate change while paving the way for more sustainable forecasting methodologies.

In conclusion, Envision's Tianji Weather Model signifies a milestone achievement in AI-powered meteorological predictions, reflecting a perfect blend of cutting-edge technology and practical application that can transform how we understand and respond to weather patterns globally.