Sub-seasonal to seasonal climate forecasting using machine learning

Abstract

[en] The main topic of this work is the study and the application of Machine Learning (ML) techniques to improve probabilistic forecasts of two-meter temperature and total precipitation at sub-seasonal scales (i.e. several weeks ahead) for the whole globe. We analyze the performance of a number of Machine Learning methods and finally we combine the best models to obtain the optimal prediction at each latitude, longitude, and for each lead time. In addition, the results of this work have been presented to an open prize challenge launched by the World Meteorological Organization (WMO) to improve current forecasts of precipitation and temperature from state-of-the-art numerical weather and climate prediction models 3 to 6 weeks into the future using Artificial Intelligence.