Hosts: Feiyu Lu and Mitch Bushuk

Machine learning techniques have seen a tremendous rise in popularity in weather and climate sciences. Data assimilation (DA), which combines observations and numerical models, has great potential to incorporate machine learning and artificial intelligence (ML/AI) techniques. One common type of modern data assimilation methods is the ensemble Kalman filter and its variants, which have been used by both research and operations in the fields of weather and climate. This study will try to apply and test some deep learning methods to compliment and improve the ensemble Kalman filter algorithm. 

The intern will work with simple (Lorenz models) and intermediate (PyQG, a Python Quasi-Geostrophic model) models to test and refine some emerging approaches that apply ML/AI methods to DA. Some recent developments include using Variational Autoencoders to generate ensemble members or using U-Net to predict ensemble covariance matrices. Given that this is an emerging and fast-developing topic, the intern will also have the opportunity to bring their own experience with machine learning and data science to data assimilation applications that could benefit future weather and climate predictions.