Existing AI models like GraphCast and Fourcastnet can already generate a weather forecast based on a set of data. But NASA and IBM have announced they have an ambitious goal to teach AI to truly forecast the weather.
NASA and IBM plan to create an AI model that will have expanded accessibility compared to existing technologies, and will also be able to process more types of data. Another key goal is to improve forecasting accuracy. It is expected that the algorithm will successfully cope with forecasting meteorological phenomena, as well as identifying conditions that contribute to the formation of various phenomena, from airplane turbulence to forest fires.
Earlier this year, NASA and IBM released another AI model that uses data from aerospace agency satellites for geospatial intelligence and is the largest open-source AI geospatial model. Since its launch, the model has been used to track and visualize tree planting and cultivation activities in Kenya. It is also used to analyze climatic conditions in the UAE.
In advancing weather and climate analysis, the joint effort between NASA and IBM promises transformative effects through innovative artificial intelligence models designed to improve accuracy and expand applications.