U-Net Kalman Filter (UNetKF): An Example of Machine Learning-assisted Ensemble Data Assimilation
This is an incremental improvement for weather and climate scientists, as it enhances data assimilation efficiency with machine learning.
The paper tackles the problem of predicting localized ensemble covariances in data assimilation for weather and climate models by using a U-Net convolutional neural network, resulting in the UNetKF method that matches or exceeds traditional methods like 3DVar and EnKF, particularly for small ensemble sizes.
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. In this paper, we use U-Net, a type of convolutional neutral network (CNN), to predict the localized ensemble covariances for the Ensemble Kalman Filter (EnKF) algorithm. Using a 2-layer quasi-geostrophic model, U-Nets are trained using data from EnKF DA experiments. The trained U-Nets are then used to predict the flow-dependent localized error covariance matrices in U-Net Kalman Filter (UNetKF) experiments, which are compared to traditional 3-dimensional variational (3DVar), ensemble 3DVar (En3DVar) and EnKF methods. The performance of UNetKF can match or exceed that of 3DVar, En3DVar or EnKF. We also demonstrate that trained U-Nets can be transferred to a higher-resolution model for UNetKF implementation, which again performs competitively to 3DVar and EnKF, particularly for small ensemble sizes.